Blog

Master of the Beat: Ben Bakson – outsmartmagazine.com


Ben Bakson grew up an only child in a small village in south Germany, watching movies, listening to music and enjoying long days on the lake with his family.  At 18, he moved to Berlin to study Economics and after graduating, he landed a marketing job at a large firm.  He found the 9 to 5 life wasn’t for that and began dabbling in music production.
His music style today is commercial with hard, emotional, and energetic beats.  He regards vocals as vital, referring to them as the soul of every song.  Music fans seem to agree as Ben Bakson has been playing to packed dancefloors every weekend for three months, spinning parties in Sao Paulo, San Francisco, Mykonos, Washington DC and NYC.
We caught up with the master of the beat for a phone chat before he travels to Houston to spin at Rebar on Saturday, Sept 24th.
What’s the worst part of doing these kind of interviews?
Ben Bakson:  Never knowing what kind of questions are going to be asked!  (Laughing) You need to have the right answer ready immediately. But I am quite quick-witted, to be honest, therefore it has not been an issue so far.
You seem like the quiet-type; someone who prefers to keep his private life to himself.
Interesting perception, because I think most people think the opposite.  They assume that I am outgoing because I’m a DJ that celebrates my life quiet publicly on social media.  They don’t realize all of that is part of my job.   I do try to keep my private life as private as possible, especially my love life.
It seems to me like a lot of deejays choose to speak their minds through their music sets.
I definitely agree, and I absolutely do that. Music is the one common language we humans have. For me, it is therapy. I can channel my emotions through music and I love when I get feedback from fans that my music helped them to get through a difficult moment.  That makes me very happy.
Would you say your music as tribal?
I’m not a fan of labeling my music as a particular genre because I honestly think my music style crosses several genres.  I feel the same way about being forced to label my sexuality.   Am I gay simply because I like sex with men?  Sex is a very small part of who I am as a person and I think everyone should have sex with the person he loves and finds attractive. I don’t think it needs an explanation, label or commenting. I’m convinced that in future the whole idea of straight, gay and whatever we do in bed won’t matter because, in the end, what counts is that we are all human and equal.
Is that why your dancefloors attract all genders and sexualities?
I try to make my dancefloor as diverse as I can. I have a lot of straight friends that I invite to my events and they love my music as much as they gay men and women. My big goal is for my sound is to be loved by everyone.  I would like to play all floors.  The big hurdle for me is that most of the big straight festivals won`t book any of the big DJs of the LGBT world.
What will it take to blur the lines between gay and straight?
We need party producers to create events that attract straight and gay people and then book DJs from both worlds that will unite all and create a unique musical journey.
Tell us about your new song, “You Got Me.”
The track was written by the talented singer OMEO. It’s about a relationship with a toxic person and how important it is to free yourself from them.
Is it a song you can relate with?
Most definitely.  I have been single for about a year now which has been the longest time in my life so far.   Honestly, I have been very happy, sharing my time with amazing friends and my beloved dog Macho. I may have met a special person but I’m not putting any stress on it or making any long-term plans.
Who is your celeb crush?
That’s easy:  JLo’s hairdresser, Chris Appleton.
If you could have any magic power, what would it be?
Sometimes I think I already have too many magic powers. Like reading people’s minds and knowing what will happen next.  I have a sixth sense!  I also think music is my magic power.    I can bring people together, make them feel happy on the dancefloor and even help them feel better at home through my podcasts. That’s pretty powerful.
Besides music, do you have a secret talent?
I think I do, but I can’t talk about it here.
What is your idea of perfect happiness?
That moment when I realize how grateful I am with all that I have.
What is your greatest fear?
The music stopping during my set.  It has happened twice so far, both due to technical issues that were out of my control.
What is your most treasured possession?
My dog, Macho.
If you could change one thing about yourself, what would it be?
There are things I’d like to change and you can be sure I’m working on changing them already.
What is the trait you most deplore in others?
Envy. I don’t really understand people who try to sabotage others. I prefer to be happy for them and support their success. It motivates me to work harder to be more successful myself.
Which living person do you most admire?
My parents. I am very thankful for what they gave me in my childhood. Education, lots of love and empathy; without them I would not be where I am now.
What do you most dislike about your appearance?
I think I am ok.  We all need to learn to relax a bit about being too self-critical and just be happy the way we are.
What makes Macho the greatest dog in the world?
You have to meet him and you will understand immediately.  The cutest, most loving, protecting, most intelligent being I have ever met.
Follow Ben on Instagram @ ben_bakson. 

source

Special master calls for help in Trump Mar-a-Lago documents fight – POLITICO

Legal
The former president also gets a deadline to produce any proof of tampering during the FBI search last month.
A car passes in front of former President Donald Trump's Mar-a-Lago resort on Feb. 11, 2022, in Palm Beach, Fla. | Joe Raedle/Getty Images
By Josh Gerstein

Link Copied
The outside expert tapped to sort through former President Donald Trump’s legal claims over documents seized from his Florida home last month is calling for backup.
U.S. District Court Judge Raymond Dearie, a New York City-based jurist named by a federal judge in Florida to act as a so-called special master in the review of more than 11,000 documents the FBI confiscated, proposed on Thursday that former Magistrate Judge James Orenstein help with the process.
“The undersigned has determined that the efficient administration of the Special Master’s duties requires the assistance of the Honorable James Orenstein,” Dearie’s proposed plan for the document review said. The order said Orenstein “has experience with complex case management, privilege review, warrant procedures, and other matters that may arise in the course of the Special Master’s duties.”
Orenstein spent 16 years as a federal magistrate in the same Brooklyn courthouse where Dearie sits. Orenstein drew attention several years ago for his role in what was semi-sarcastically dubbed “the magistrates’ revolt” — rulings from a smattering of federal magistrate judges across the country questioning government tactics in warrant applications seeking electronic data.
In 2016, Orenstein issued a controversial ruling rejecting prosecutors’ arguments that a two-century-old federal law gave the government the right to command Apple to assist in unlocking an iPhone used by an alleged drug dealer. The judge’s pro-privacy stance in that matter may have led the secretive Foreign Intelligence Surveillance Court to appoint him to a list of approved friends of the court who provide their perspective on surveillance requests.
Dearie proposed that the former magistrate, who has a top secret clearance, be paid $500 an hour for work on the Trump documents case. The federal judge who named Dearie as special master, Florida-based Aileen Cannon, has previously ruled that Trump must assume all expenses related to the review.
Dearie said in the Thursday order that he would not be compensated for his work because he’s still being paid his judge’s salary, but that expenses such as Orenstein’s fees must be paid in full and on time or Trump could face court sanctions.
While the seven-page proposal the special master issued on Thursday bears Dearie’s name and a form of electronic signature, metadata attached to the document indicates that Orenstein — who retired from the court in 2020 — was involved in drafting it. Dearie’s chambers did not immediately respond to a request for comment on whether Orenstein helped prepare the proposal to involve him in the case.
Dearie’s plan, still subject to comment from the parties, gives Trump a deadline of Sept. 30 to indicate whether he disputes the FBI’s list of what was seized in the search. Last month, Trump suggested that such tampering was a possibility given the FBI’s refusal to allow a lawyer for Trump to witness the search.
“The FBI and others from the Federal Government would not let anyone, including my lawyers, be anywhere near the areas that were rummaged and otherwise looked at during the raid on Mar-a-Lago,” Trump wrote on his Truth Social media site last month. “Everyone was asked to leave the premises, they wanted to be left alone, without any witnesses to see what they were doing, taking or, hopefully not, ‘planting.’”
Trump has not offered any evidence of such tampering by the FBI, but Dearie’s plan says that if Trump’s team wants to press such a claim, someone will need to submit a statement under penalty of perjury detailing the grounds for it.
The proposed plan calls for the review of the more than 11,000 documents seized from Mar-a-Lago to take place in four batches. The first involves a small set of information the Justice Department has already flagged as potentially subject to attorney-client privilege. Trump’s team has weekly deadlines stretching through Oct. 14 to identify allegedly privileged material and explain why. Prosecutors have parallel deadlines through Oct. 21, with Dearie scheduled to submit his recommendations to Cannon by Oct. 31
As of now, roughly 100 documents with various national security classification markings that were seized during the search will not be part of the special master process. Cannon ruled that they should be and rebuffed a request by prosecutors to carve those records out.
But on Wednesday night, a federal appeals court sided with the Justice Department, ruling that the government’s broad authority over classified information made it a mistake for Cannon, a Trump appointee, to include the alleged secrets in the wider review and for her to put them off-limits to criminal investigators in the meantime.
Trump’s lawyers have yet to indicate whether they plan to ask the Supreme Court to step in and restore Cannon’s original order.
It took little time for the 11th U.S. Circuit Court of Appeals ruling against Trump to boomerang and help prompt a legal loss for one of his allies: MyPillow founder Mike Lindell.
Last week, the FBI used a search warrant to seize Lindell’s mobile phone as he waited at a Hardee’s drive-through in Mankato, Minn. Lindell filed suit on Tuesday in federal court in Minnesota, seeking a temporary restraining order requiring the government to stop accessing his phone and ultimately to return it.
However, a judge denied the restraining order on Thursday, citing, among other things, the 11th Circuit ruling against Trump a day earlier.
U.S. District Court Judge Eric Tostrud, a Trump appointee, said that Lindell’s right to get his phone back “is not obvious, and that’s understating things.” The judge’s order didn’t rule out the possibility that Lindell might be entitled to get his phone back after both sides have the chance to develop their legal arguments in court filings.
Lindell’s phone appears to have been seized in connection with an ongoing investigation into alleged tampering with voting machines in Colorado last year.
CORRECTION: An earlier version of this article misstated the focus of the investigation that led to the seizure of Mike Lindell’s phone.
Link Copied
© 2022 POLITICO LLC

source

Coventry University to name building after electronic pioneer Delia Derbyshire – Mixmag

Coventry University to name building after electronic pioneer Delia Derbyshire  Mixmag
source

Howl's Moving Castle – Studio Ghibli Fest 2022 – Creative Loafing Tampa

Rated PG

The next date Howl’s Moving Castle – Studio Ghibli Fest 2022 is playing is Sunday, September 25 .
(Howl’s Moving Castle – Studio Ghibli Fest 2022 is not showing in any area theaters on Saturday, September 24.)
Subscribe now to get the latest news delivered right to your inbox.
Tampa now has a magic mushroom dispensary
By Colin Wolf
Family of Andrew Joseph III awarded $15 million in lawsuit against Hillsborough Sheriff, calls win a blow to qualified immunity
By Justin Garcia
Florida Cracker Fish Company will open out of former Ballyhoo Grill location in Citrus Park
By Kyla Fields
St. Pete’s Independent Bar is closing, new brewery in Seminole Heights, and more in local foodie news
By Kyla Fields
View more issues
Read our sister publications
633 N. Franklin Street, Suite 735
Tampa, FL 33602

source

A high-resolution flux-matrix model describes the spread of diseases in a spatial network and the effect of mitigation strategies | Scientific Reports – Nature.com

Thank you for visiting nature.com. You are using a browser version with limited support for CSS. To obtain the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles and JavaScript.
Advertisement
Scientific Reports volume 12, Article number: 15946 (2022)
Metrics details
Propagation of an epidemic across a spatial network of communities is described by a variant of the SIR model accompanied by an intercommunity infectivity matrix. This matrix is estimated from fluxes between communities, obtained from cell-phone tracking data recorded in the USA between March 2020 and February 2021. We apply this model to the SARS-CoV-2 pandemic by fitting just one global parameter representing the frequency of interaction between individuals. We find that the predicted infections agree reasonably well with the reported cases. We clearly see the effect of “shelter-in-place” policies introduced at the onset of the pandemic. Interestingly, a model with uniform transmission rates produces similar results, suggesting that the epidemic transmission was deeply influenced by air travel. We then study the effect of alternative mitigation policies, in particular restricting long-range travel. We find that this policy is successful in decreasing the epidemic size and slowing down the spread, but less effective than the shelter-in-place policy. This policy can result in a pulled wave of infections. We express its velocity and characterize the shape of the traveling front as a function of the epidemiological parameters. Finally, we discuss a policy of selectively constraining travel based on an edge-betweenness criterion.
When plague hit Florence in August 1630, the Florentine authorities made a number of high-stakes decisions which proved highly effective1. One of the reasons the Florence Sanitá could organize this response was the ample time they had, forewarned as they were by the Milanese authorities in November 1629. Today’s public-health authorities work under much more compressed timescales, as evidenced by the SARS-CoV-2 pandemic. Long-distance travel radically changes the dynamics of spreading, which raises a number of questions about the spatial dynamics of transmission in modern times. Epidemic outbreaks in the last two decades have provided the scientific community with a wealth of material to study these questions, going beyond the classic Susceptible-Infected-Recovered (SIR) theory with perfect mixing2,3,4,5,6. Several studies have shown how the total epidemic size can be affected by factors such as inhomogeneity in transmission rates7,8,9,10,11,12,13,14 or in the mode of transmission15,16. Classically, motion of individuals was taken into account by introducing diffusion terms in the standard SIR equations, allowing the emergence of spatio-temporal patterns17,18,19,20. Recently, in the context of the SARS-CoV-2 pandemic, such approaches have been especially valuable in order to study the effect of containment policies such as lockdown and quarantine21,22. These models are, however, limited in that they do not, in principle, take long-distance air travel into account. Several works have, therefore, considered disease spread in a network, typically constructed from air-traffic data19,23,24, where edges can connect locations separated by large geographical distances. This approach can lead to very accurate predictions at the country scale25 but predictions at finer scales remain challenging. Another study considering human mobility emphasized how spatial variation in public-health infrastructure reflected on epidemiological parameters can affect the dynamics of spread to different countries26. Data-based studies of epidemic spread and the impact of social distancing through the analysis of social-network structure have also been very informative16,27,28,29.
A recent paper by Chang et al.30 obtained a model for the spatio-temporal spread of a disease at a high spatial resolution by using extensive mobile tracking information to identify physical interactions between individuals. Chang et al. showed that the actual spread of the SARS-CoV-2 epidemic can be well explained from the mobility data of individuals. The model relied on the simulation of interactions among individuals on a bipartite graph where nodes, representing locations at a very fine spatial resolution, are divided in two sets: Census Block Groups (residential areas) and Points Of Interest (non-residential), each of them having its own transmission rate. The model was fitted to reproduce known reported cases of COVID-19 in 10 metropolitan areas, and could then be used to make short-term predictions about the spreading or study the effect of different mitigation strategies.
Here we take an approach similar to that of Chang et al., using mobility data to calibrate a model for disease spread. However, we investigate this propagation at the scale of a large country, the USA, rather than metropolitan areas. Specifically, we introduce a spatial SIR epidemiological model in which effective transmission rates between (N={2^{10}}) communities are computed from mobility data of individuals belonging to these communities. We show that this model captures well the spread of the SARS-CoV-2 epidemic. Remarkably, we find that a simple model consisting of an interaction frequency dropping under the effect of lockdown, and of a single flux matrix encoding the travel of individuals, faithfully reproduces the reported cases of COVID-19 both globally and locally in each community. Strikingly, the SARS-CoV-2 epidemic spreads in a delocalized fashion, infecting distant communities very quickly. Moreover, an even simpler model with uniform transmission rates between the communities gives results very close to the model based on mobility data, emphasizing the prevalent role air travel played in the spread of the SARS-CoV-2 epidemic. We then study how interventions that change travel patterns can localize epidemics. In particular, we investigate the hypothetical effect of a policy preventing long-distance travel. In addition to “flattening” the curve, spreading through nearest-neighbor interactions creates traveling waves, which we characterize both analytically and numerically. These results allow us to discuss which interventions are more effective, limiting short-range contacts (a lockdown), or limiting long-range trips (a quarantine). We also propose an alternative mitigation strategy based on an edge-betweenness criterion.
We consider a metapopulation model of N communities numbered 1, 2,…, N. Denoting by (S_a), (I_a) and (R_a) the numbers of susceptible, infected and recovered individuals in community a, the standard SIR equations read:
where (beta _{ab}) is the transmission rate from infected individuals in community b to susceptible individuals in community a and (gamma) is the recovery rate, assumed to be the same for all communities. Diagonal elements of the infectivity matrix ([beta _{ab}]) describe intracommunity infections, while off-diagonal elements describe inter-community infections (Fig. 1a,b). We also introduce the local epidemic sizes (T_a = I_a + R_a). The total population in each community (M_a) is constant through time,
The model in Eq. (1) has been extensively studied31,32,33,34,35. We show in the Supplementary Information (SI) that the dynamics can be reduced to an ODE of just one N-vector variable, and the endemic equilibrium can be obtained by solving a transcendental equation involving the infectivity matrix ([beta _{ab}]).
Model for the spreading of SARS-CoV-2 in a network of communities in the USA. (a) One community i interacting with three other communities j, k and l, with the transmission rates (beta _{ij}), (beta _{ik}) and (beta _{il}) respectively. (b) Infectivity matrix. (c) (N=2^{10}) communities in the USA. Each community aggregates a number of Census Block Groups (CBGs).
In order to estimate the infectivity matrix, we used mobility data compiled by SafeGraph36, tracking the location of about 20 million USA cell phones between March 2020 and February 2021. The locations consist of more than 200,000 Census Block Groups (CBGs). Each cell phone is assigned for physical residence the location where it spent the most time, and daily visits to other locations are recorded. For computational purposes, we coarse-grained the physical locations into (N = {2^{10}}) communities, which are shown in Fig. 1c. Let (f_{ab}) be the number of individuals from community a visiting community b per unit of time. We will assume that
The variation in susceptible individuals in community a due to new infections during the time interval (Delta t) has the form:
where (beta) is the disease-specific transmission rate when a susceptible individual has contact with an infected individual. There are three kinds of infection to consider: (i) the infected person and the infector belong to the same community, (ii) an infected person visits a neighboring community and infects a resident of this community, and (iii) a susceptible person visits a neighboring community and gets infected by one of its residents. We will neglect the rarer “tourist to tourist” infections, when an infected person visits a neighboring community and infects there a visitor from yet another community. The term (mathrm {Pr}left( text {meeting an infected individual}right)) can therefore be evaluated as a function of the pseudo-flux matrix ([f_{ab}]) for each of the three aforementioned cases (SI). After summation of the three contributions, we obtain:
where p is the frequency with which an individual is having contact with an other individual of its community, and is assumed to be the same for all communities. Equation (5) determines the infectivity matrix ([beta _{ab}]) from Eq. (1) up to a proportionality constant, namely (beta p).
In order to assess the validity of the infectivity matrix based on mobility data, Eq. (5), we confronted the model’s predictions against COVID-19 case numbers reported in the USA by the Center for Systems Science and Engineering (CSSE) at Johns Hopkins University37. Specifically, we fitted the daily (beta p(t)) values so as to minimize the sum of squared errors between values predicted by the model and values reported by the CSSE (Material and methods). The fitted (beta p(t)) values show a steep decay during the month of March 2020, followed by a plateau lasting until February 2021 (Fig. 2b).
Model based on SafeGraph mobility data or a uniform infectivity matrix. (a) The model is fitted to COVID-19 confirmed cases in the USA. There is one fitting parameter per day. (b) Fitting parameters obtained. The shape suggests a simplified model with two limiting values before and after lockdown. (c) The simplified models obtained reproduce the spread of the SARS-CoV-2 epidemic in the community network. A direct comparison between the local epidemic sizes predicted by the models and the reported values can be found in Fig. 5. See also Movies S1 and S2.
The time variations of (beta p(t)) shown in Fig. 2b imply a drastic decrease in the interaction frequency among individuals across all communities. This result reflects the effect of the shelter-in-place policies that were implemented at the beginning of the SARS-CoV-2 pandemic in many USA states. We can use this observation to estimate the effect of these policies quantitatively: the interaction frequency among individuals, namely p, is about 5 times smaller in the plateau following shelter-in-place policies than it was at the onset of the SARS-CoV-2 pandemic. Following this observation, we defined a simplified model, in which the fit with reported cases was carried out while enforcing a softplus shape for (beta p(t)) (Fig. 2b and Material and methods). This simplified model reproduced the average progression of the epidemic (Fig. 2a), but it didn’t capture the three oscillations visible in the number of new cases. We conclude that these oscillations in the number of new cases were mostly driven by a similar oscillatory pattern in the interaction frequency as seen in Fig. 2b. In Fig. 2c, we show for a few dates an overlay of the number of new COVID-19 cases in each community according to the official reports and as predicted by our simplified model. Despite non-negligeable variance, we find that the model reproduces to a satisfying degree the real dynamics. This can also be seen through a direct comparison of the model predictions with the real reports of new cases, for all dates combined, as shown in Fig. 5a. This result suggests that the infectivity matrix constructed from mobility data (Eq. 5) is a reasonable approximation of the “true” infectivity matrix.
The previous results suggest that the decrease in new COVID-19 cases was mostly driven by a country-wide reduction in the interaction frequency among individuals. Here we investigate the hypothetical effect of an alternative policy, namely a travel restriction while keeping unchanged the interaction frequency among individuals. Specifically, we modified the infectivity matrix so that communities separated by a physical distance larger than a prescribed cutoff do not interact: (beta _{ij} = 0) if (d_{ij} > d_c) (Fig. 3a). We seeded the infection in a community belonging to the state of Washington and simulated the spread of the epidemics using a fixed interaction frequency ((beta p(0)) from the simplified model). As expected, we observed a reduction in the number of daily new cases dT (we define the local variables (dT_a(t) = T_a(t)-T_a(t-1))) when the cutoff distance (d_c) decreased, illustrating the “curve-flattening” effect that was targeted by travel restriction policies (Fig. 3b). In this idealized scenario with a single seed for the infection, the epidemic propagates as a traveling wave from the west coast to the east coast (Fig. 3c). However, as can be seen by comparing Fig. 2a to Fig. 3b, travel restriction policies are not as efficient as lockdown policies to decrease the spread of an epidemic.
Limiting long-distance travels without local lockdown. (a) Only transmission rates for communities separated by a distance (d_{ij} < d_c) are retained. (b) Daily new infections for increasing values of the cutoff distance. (c) Spatial visualization of the daily new infections using a cutoff (d_c = {200};{text {km}}). See also Movie S3.
The daily infectivity matrices constructed from the SafeGraph mobility data can be viewed as elements of a random-matrix ensemble. Remarkably, matrix elements seem to be distributed according to the law (beta _{ij} = langle beta _{ij} rangle e^{xi }), where (langle beta _{ij} rangle) is the mean infectivity matrix and (xi equiv N(0,1)) is a centered reduced Gaussian variable (Fig. S1). The probability density of the eigenvalues is also shown in Fig. S1. A connection can be made with random matrix theory (RMT)38,39,40, initially introduced by E. Wigner to model the spectra of the nuclei of heavy atoms, where the interactions between many nucleons are assumed to be drawn from a random ensemble. In RMT, random matrices are classified according to their symmetry, or according to their corresponding level statistics (namely, the probability density of the spacing between consecutive eigenvalues) that exhibit different degrees of level repulsion38. In particular, the Wigner-Dyson (WD) statistics is typical of the Gaussian Orthogonal Ensemble (in which eigenvalues “interact”), while the Poisson statistics is typical of a matrix with independent eigenvalues. The level-spacing statistics of the infectivity matrices is shown in Fig. 4a. We found that they were time-independent. Yet interestingly, the level statistics interpolates between the WD statistics and the Poisson statistics41. As shown earlier, removing links between communities according to their geographical distance results in a slower spread and a smaller epidemic size. Concomitantly, the level spacing distribution converges toward a Poisson statistics (Fig. 4c). The crossover from WD (entropy (S=0.7169)) to Poisson ((S=1)) distribution as links between communities are successively removed suggests an isolation policy that can lead to an effective reduction of the epidemic size (SI). As alternative mitigation strategies, we choose to induce the transition toward a Poisson distribution by decimating links between communities according to their “nominal” distance or the “edge-betweenness” centrality42,43,44 (Fig. 4b). Figure 4d shows how the level spacing distribution converges toward a Poisson statistics when the nominal distance threshold is lowered. We find the edge-betweenness centrality to be more efficient in decreasing the epidemic size. This is because decimation according to edge-betweenness first targets links with the largest transmission rates. One could also consider a moderate policy: instead of eliminating links completely, one could impose constraints on the flux of individuals that are allowed to commute via central pre-determined links.
Random-matrix properties and mitigation strategies. (a) Level spacing distribution (unfolded38,45,46,47) of infectivity matrices between 2020-03-01 and 2021-02-15 (pooled by 7-day windows). The solid lines denote the Wigner-Dyson and the Poisson statistics. The level distribution is shown in Fig. S1. (b) The epidemic size (Omega /Omega _0) versus the relative number of cuts (N_c/N/(N-1)) for the 3 mitigation methods: distance cutoff, bandwidth, and edge-betweenness centrality. (Omega _0) is the epidemic size for (N_c=0). Level-spacing entropy S and epidemic size as a function of (c) the cutoff distance (d_c) in the range 170-12,000 km, and (d) the half-bandwidth (1{le }B<1023).
Although SafeGraph mobility data36 provide a realistic picture of people movement between communities, one might ask what is the sensitivity of the results to the infectivity matrix derived from the mobility fluxes. We thus considered a model with uniform transmission rates among communities, namely (beta _{ab} = beta), which suppresses spatial effects. In particular, this model leads to the natural variables (nu _a) (see SI) to be uniform: (nu _a(t) = nu ^text {UN}(t)). Surprisingly, carrying out the same fit to the reported cases (Fig. 1a,b) was only marginally inferior to the fit carried out with the SafeGraph mobility data (see Fig. 5b). By contrast, carrying out the same fit with the infectivity matrix derived from SafeGraph mobility data but with long-range interactions turned down resulted in a significantly different dynamics (see Fig. 5c). There are several implications of this result. (1) Although changes in the structure of the infectivity matrix can lead to drastically different dynamics (Figs. 3, 4 and 5c), it appears that the infectivity matrix derived from SafeGraph fluxes falls in the same “universality class” as the uniform model. This suggests that long-range movements (e.g. air traffic) played a prevalent role in the spread of SARS-CoV-2. (2) Our choice to reduce the complexity of the model to only one fitting parameter might not be adequate to discriminate between models falling into the same universality class. Instead of fitting only one scalar p(t) at each time point, we also investigated the possibility of fitting the (N^2) transmission rates (beta _{ab}(t)) minimizing the errors with reported cases (see Material and methods and Fig. S2). Although this latter approach is clearly prone to overfitting, it shows that there exists a parametrization of the model which reproduces very closely reported cases. We anticipate that the proper number of fitting parameters lies in between those two extreme scenarios.
Comparison of the (fitted) model predictions with reported values using (a) the infectivity matrix derived from SafeGraph fluxes (same model as Fig. 2), (b) an infectivity matrix with uniform transmission rates, and (c) the infectivity matrix derived from SafeGraph fluxes truncated to geographical distances (d_c < {400} ; text{km}). Local epidemic sizes (T_a) are shown. One symbol is associated to one given community and one given day.
The results from the previous sections suggest the apparition of a wavefront when transmissions are short range. To investigate this phenomenon, we consider a simplified model of communities lying on a two-dimensional square lattice. Each community index is replaced by coordinates ((i,j) in llbracket 1, 2^n rrbracket times llbracket 1, 2^m rrbracket). In particular, we consider that only individuals from neighboring communities interact together. We rewrite Eq. (1):
where (alpha) (respectively (beta)) is the intra-community (resp. inter-community) transmission rate. After rescaling the time and space variables, and defining the rescaled recovery rate (tilde{gamma } = gamma / (a beta )) where (a = 4 + alpha /beta), we look for wave solutions in the continuum limit by introducing the shape functions (S(x,y,t) = g(x – tilde{v} t)) and (I(x,y,t) = h(x – tilde{v} t)), where (tilde{v} = v / sqrt{a} beta) represents the velocity of the wave in rescaled time and space (see SI). The shape functions satisfy the ODE:
We find that the velocity of the traveling wave is bounded from below (SI):
which is in agreement with previous reports18,48 and with results from marginal-stability analysis21,49,50. Interestingly, by an independant argument, we also established (see SI) that (tilde{v} le tilde{v}_c). Therefore the traveling wave must move at velocity (tilde{v} = tilde{v}_c). This indicates that the SIR dynamics in Eq. (6) falls into the Fisher-Kolmogorov-Petrovsky-Piscunov universality class, resulting in pulled waves51,52,53,54. Although we have taken the continuum limit of a nearest-neighbor model, this analysis is also valid for any finite-range infection matrix with the appropriate rescaling of variables.
We performed simulations of the dynamics given by Eq. (6) on a square lattice with a varying aspect ratio (Fig. 6a) and seeding the infection at one site on the west boundary. As expected, there is a front of new infections, moving from west to east as time progresses. A timelapse of a traveling front of infected individuals with (alpha = beta = gamma = {0.1}) is shown in Fig. 6b. The wave position increases asymptotically linearly with time, but the velocity (tilde{v}) of the wave varies with (tilde{gamma }) (Fig. 6c). The profiles obtained are in agreement with the shape functions obtained by solving Eq. (7), as shown in Fig. 6d.
Existence of a wave with nearest-neighbors-only interactions. (a) Equation (6) is solved on a square lattice of (2^n times 2^m) sites. (b) Traveling front of infected individuals moving along the x direction (left to right). We took (alpha = beta = gamma = {0.1}), so that the rescaled recovery rate is (tilde{gamma } = gamma / (a beta ) = {0.2}). (c) The dynamics simulated for different values of (beta). The position m(t) ((S(t, m(t))=(1 + S_infty )/2)) of the wave is asymptotically linear in time, hence a constant velocity (tilde{v} = sqrt{1-tilde{gamma }}). (d) The simulated wave profiles (symbols) are in agreement with the profiles predicted by Eq. (7) (solid lines). The corresponding S,I,R profiles are shown in Fig. S3. A simulation with (tilde{gamma } = {0.2}) is shown in Movie S4.
Biological systems are inherently complex and it can be a challenge to characterize them by a small number of parameters. In the case of epidemics, the huge number of parameters (for example, intercommunity spread involves (N^2) transmission rates for N communities) can obscure the salient mechanisms and make it difficult for policy makers to find efficient interventions. Reducing the dimension of a model, when possible, is therefore of great value. In this article, we have proposed a spatial SIR model with an infectivity matrix based on the local travel patterns.
Despite its simplicity (there is only one global parameter to be adjusted), we find that our model is able to capture the spatial spread of the SARS-CoV-2 epidemic, with a delocalized multicenter spreading caused by long-distance travelers bringing infection into far regions, then becoming secondary centers of infection. The time evolution of the interaction frequency reflects the shelter-in-place policies that were implemented by various USA states in the early stages of the SARS-CoV-2 pandemic. In fact, the rich diversity of human responses could be summarized by a simplified model for the interaction frequency, whose asymptotes represent the values “before lockdown” and “after lockdown”, with few assumptions about the infection process.
If a model with relatively few parameters describes the observations, one can assume that it also describes the situation when these parameters are changed by an intervention. Therefore, we suggest such a model could be used as the basis for efficient policies—or at least reliably to estimate the consequences of adopting policies. As an example, we have shown the hypothetical effect of an alternative to the shelter-in-place policy in the case of the SARS-CoV-2 pandemic. Specifically, we have investigated a travel restriction policy in which individuals can only move within an area of fixed radius centered around their residence. By contrast to the lockdown policy, we find that infections spread through a well-defined wave front, traveling with a certain velocity. This scenario might be preferable since it gives time to communities and public health infrastructures to prepare for the onset of the epidemics, while still “flattening the curve” of new infections (Fig. 3b). We have also provided both an analytical and numerical analysis elucidating the mechanism of formation of a wavefront. In particular, we give the velocity and the shape of the wavefront for an epidemic spreading through nearest neighbors interactions.
During the course of our research, an epidemiological study bearing similarities with our approach was published55. In that study, the authors developed a county-resolved metapopulation model describing the spread of SARS-CoV-2 in the USA, informed with mobility data from SafeGraph. Instead of calibrating the model by fitting the country-wide number of reported cases as in the present study, the authors fitted their model to reproduce reported cases of COVID-19 on a per-county basis. Furthermore, the model required fitting many time-dependent parameters on a per-county basis, including per-county transmission rates, making the fitting procedure very high dimensional. Finally, the dynamics of the disease spread introduced differs from Eq. (1) since S, I and R compartments were introduced for each commute channel (i rightarrow j) rather than for each community. Altogether, the complexity of that model makes it less amenable to analytical study.
In conclusion, we used a simple model of intercommunity spread of an infectious disease to show the transition between different regimes of epidemic progression. Because of the complexity of the infection process (e.g., variations in individuals’ responses, mutations, etc.) and of human behavior, we are still far from a global forecasting system able to predict the spread of different infectious agents throughout the world. As with weather forecasting, observables must be measured in real time in order to inform complex models to yield short-term forecasts. One salient feature in our approach is showing how such widespread measurements (namely, the mobility data) can be integrated in a model describing the spread of an infectious agent, and showing what types of predictions can be obtained. It is a step toward more predictive epidemiology models, grounded in measurable quantities.
The mobility data was obtained from SafeGraph, a company that aggregates anonymized location data from numerous applications in order to provide insights about physical places, via the SafeGraph Community. To enhance privacy, SafeGraph excludes census block group (CBG) information if fewer than two devices visited an establishment in a month from a given census block group. In this manuscript, we use data extracted from the “Social Distancing Metrics” dataset36, with dates between February 2020 and February 2021. SafeGraph has stopped sharing mobility data under this format, however the same data can be acessed under a different format through the Neighborhood Patterns dataset56. However the matrices of fluxes between our coarse-grained communities can be found in the ‘data’ folder in our github repository. We obtained the coarse-grained communities by running a K-means clustering algorithm to group the 220,333 CBGs into ({2^{10}}) communities. We used the implementation from Scikit Learn. We then ran a hierarchical clustering algorithm to re-index communities so that communities close in space had close indices, as shown in Fig. 1. We used the linkage function from SciPy. We then computed for every day the “flux matrix” where each entry (f_{ab}) represents the number of cell phone whose residence CBG belongs to community a which visited a CBG belonging to community b. The average flux matrix was constructed by averaging all the daily flux matrices. We used population counts for CBGs in agreement with the United States Census Bureau’s and available in the SafeGraph Open Census Data57 (file “cbg_b01.csv”, column “B01001e1”). We checked that the population counts from the United States Census Bureau were approximately proportional to the residential mobile-phone counts, therefore validating mobile tracking as a proxy for actual population.
In order to fit our model, we used USA cases of COVID-19 reported by the Center for Systems Science and Engineering (CSSE) at Johns Hopkins University37 which can be accessed at the GitHub repository github.com/CSSEGISandData/COVID-19. New cases of COVID-19 reported by the CSSE were mapped to the closest mobility-data-derived community based on their latitude and longitude. We therefore obtained the time evolution of SARS-CoV-2 infections through the mobility-data derived communities. Integer absolute new cases were converted into relative population fractions using the community population counts obtained from the United States Census Bureau.
Here we consider the SIR dynamics (see Supplementary Information):
We also define the chi-square at each time t:
where (M_a) is the population at location a and ((1 – hat{s}_a) M_a) is the number of reported cases at location a. Assuming that (p(t-1), p(t-2), ldots , p(0)) have been previously evaluated, the dynamics ((s_a, j_a)) is determined up to time t, and we have:
To obtain the scale p(t), we solve for:
We look for a simplified model in which the scales have a functional form close to a ramp function:
We first set (theta) by fitting this function to the daily scales:
Then we adjust the slope of the ramp in order to minimize the error with reported cases. In particular, we set (theta = (theta _1^*, theta _2^*, psi , theta _4^*)), and we solve:
We thus define (theta ^text {simplified} = (theta _1^*, theta _2^*, psi ^*, theta _4^*)), and the model for simplified scales is:
We also considered another approach, consisting of fitting the (N^2) transmission rates (beta _{ab}(t)) at every time t. In particular, we solve:
subject to the constraints (beta _{ab} ge 0). We show the results of this fit in Fig. S2, however this approach is prone to overfitting since there are (N^2) fitting parameters and only N data point at each time t.
The curves shown in Fig. 6b,c and the symbols shown in Fig. 6d were obtained by integrating Eq. (6).We used the function solve_ivp from SciPy with the “DOP853” integration method. At (t = 0), we considered (S = 1) everywhere except at the sites of coordinates ((0, 2^{m-1}-1)) and ((0, 2^{m-1})) (see Fig. 6a), where we set (S=0) and (I = 1). The Laplacian was computed using the 9-point stencil (Delta _text {discrete} = left( {begin{matrix} 1 &{} 2 &{} 1 \ 2 &{} -12 &{} 2 \ 1 &{} 2 &{} 1 end{matrix}}right) /4). We considered periodic boundary conditions along the y direction and Dirichlet boundary conditions along the x direction.
Data and scripts used in this study are available at the GitHub repository github.com/czbiohub/epidemiology_flux_model. Except the raw mobility data, which belongs to SafeGraph.
Henderson, J. Florence Under Siege: Surviving Plague in an Early Modern City (Yale University Press, 2019).
Google Scholar 
Kermack, W. O. & McKendrick, A. G. A contribution to the mathematical theory of epidemics. Proc. R. Soc. Lond. Ser. A Contain. Pap. Math. Phys. Charact. 115, 700–721 (1927).
ADS  MATH  Google Scholar 
Hethcote, H. W. The mathematics of infectious diseases. SIAM Rev. 42, 599–653. https://doi.org/10.1137/S0036144500371907 (2020).
ADS  MathSciNet  MATH  Google Scholar 
Harko, T., Lobo, F. S. & Mak, M. Exact analytical solutions of the susceptible-infected-recovered (SIR) epidemic model and of the SIR model with equal death and birth rates. Appl. Math. Comput. 236, 184–194 (2014).
MathSciNet  MATH  Google Scholar 
Wu, J. T., Leung, K. & Leung, G. M. Nowcasting and forecasting the potential domestic and international spread of the 2019-nCoV outbreak originating in Wuhan, China: A modelling study. Lancet 395, 689–697. https://doi.org/10.1016/S0140-6736(20)30260-9 (2020).
CAS  PubMed  PubMed Central  Google Scholar 
Kucharski, A. J. et al. Early dynamics of transmission and control of COVID-19: A mathematical modelling study. Lancet Infect. Dis. 20, 535–558. https://doi.org/10.1016/S1473-3099(20)30144-4 (2020).
Google Scholar 
Allard, A., Moore, C., Scarpino, S. V., Althouse, B. M. & Hébert-Dufresne, L. The role of directionality, heterogeneity and correlations in epidemic risk and spread (2020). arXiv preprint arXiv:2005.11283.
Aleta, A. et al. Quantifying the importance and location of SARS-CoV-2 transmission events in large metropolitan areas. medRxiv https://doi.org/10.1101/2020.12.15.20248273 (2020).
PubMed  PubMed Central  Google Scholar 
Hébert-Dufresne, L., Althouse, B. M., Scarpino, S. V. & Allard, A. Beyond (R_0): Heterogeneity in secondary infections and probabilistic epidemic forecasting. J. R. Soc. Interface 17, 20200393 (2020).
PubMed  PubMed Central  Google Scholar 
Britton, T., Ball, F. & Trapman, P. A mathematical model reveals the influence of population heterogeneity on herd immunity to SARS-CoV-2. Science 369, 846–849 (2020).
ADS  MathSciNet  CAS  PubMed  PubMed Central  MATH  Google Scholar 
Neipel, J., Bauermann, J., Bo, S., Harmon, T. & Jülicher, F. Power-law population heterogeneity governs epidemic waves. PLoS ONE 15, e0239678 (2020).
CAS  PubMed  PubMed Central  Google Scholar 
Sun, K. et al. Transmission heterogeneities, kinetics, and controllability of SARS-CoV-2. Science https://doi.org/10.1126/science.abe2424 (2021).
PubMed  PubMed Central  Google Scholar 
Kawagoe, K. et al. Epidemic dynamics in inhomogeneous populations and the role of superspreaders. Phys. Rev. Res. 3, 033283. https://doi.org/10.1103/PhysRevResearch.3.033283 (2021).
CAS  Google Scholar 
Pozderac, C. & Skinner, B. Superspreading of SARS-CoV-2 in the USA. PLoS ONE 16, 1–10. https://doi.org/10.1371/journal.pone.0248808 (2021).
CAS  Google Scholar 
Huber, G. et al. A minimal model for household effects in epidemics. Phys. Biol. 17, 065010 (2020).
PubMed  Google Scholar 
Aleta, A. et al. Modelling the impact of testing, contact tracing and household quarantine on second waves of COVID-19. Nat. Hum. Behav. 4, 964–971. https://doi.org/10.1038/s41562-020-0931-9 (2020).
PubMed  PubMed Central  Google Scholar 
Murray, J. Mathematical Biology II. Spatial Models and Biological Applications (Springer, 2003).
Google Scholar 
Postnikov, E. B. & Sokolov, I. M. Continuum description of a contact infection spread in a SIR model. Math. Biosci. 208, 205–215 (2007).
MathSciNet  PubMed  MATH  Google Scholar 
Brockmann, D. & Helbing, D. The hidden geometry of complex, network-driven contagion phenomena. Science 342, 1337–1342 (2013).
ADS  CAS  PubMed  Google Scholar 
Chu, A., Huber, G., McGeever, A., Veytsman, B. & Yllanes, D. A random-walk-based epidemiological model. Sci. Rep. 11, 19308. https://doi.org/10.1038/s41598-021-98211-5 (2021).
ADS  CAS  PubMed  PubMed Central  Google Scholar 
Te Vrugt, M., Bickmann, J. & Wittkowski, R. Effects of social distancing and isolation on epidemic spreading modeled via dynamical density functional theory. Nat. Commun. 11, 1–11 (2020).
Google Scholar 
Tsori, Y. & Granek, R. Epidemiological model for the inhomogeneous spatial spreading of COVID-19 and other diseases. PLoS ONE 16, e0246056 (2021).
CAS  PubMed  PubMed Central  Google Scholar 
Tizzoni, M. et al. Real-time numerical forecast of global epidemic spreading: case study of 2009 A/H1N1pdm. BMC Med. 10, 1–31 (2012).
Google Scholar 
Linka, K., Peirlinck, M., Sahli Costabal, F. & Kuhl, E. Outbreak dynamics of COVID-19 in Europe and the effect of travel restrictions. Comput. Methods Biomech. Biomed. Eng. 23, 710–717 (2020).
MATH  Google Scholar 
Ivorra, B., Ngom, D. & Ramos, A. M. Be-CoDiS: A mathematical model to predict the risk of human diseases spread between countries. validation and application to the 2014-15 Ebola virus disease epidemic. arXiv preprint arXiv:1410.6153 (2014).
Hsu, S. & Zee, A. Global spread of infectious diseases. J. Biol. Syst. 12, 289–300 (2004).
MATH  Google Scholar 
Nande, A., Adlam, B., Sheen, J., Levy, M. Z. & Hill, A. L. Dynamics of COVID-19 under social distancing measures are driven by transmission network structure. PLoS Comput. Biol. 17, e1008684 https://doi.org/10.1371/journal.pcbi.1008684 (2021).
ADS  CAS  PubMed  PubMed Central  Google Scholar 
Ventura, P. C., Aleta, A., Rodrigues, F. A. & Moreno, Y. Modeling the effects of social distancing on the large-scale spreading of diseases. arXiv:2105.09697 (2021).
Mayberry, J., Nattestad, M. & Tuttle, A. The structure of an outbreak on a college campus. Math. Mag. 94, 83–98 (2021).
MathSciNet  MATH  Google Scholar 
Chang, S. et al. Mobility network models of COVID-19 explain inequities and inform reopening. Nature 589, 82–87 (2021).
ADS  CAS  PubMed  Google Scholar 
Hethcote, H. W. An immunization model for a heterogeneous population. Theor. Popul. Biol. 14, 338–349 (1978).
MathSciNet  CAS  PubMed  MATH  Google Scholar 
Post, W., DeAngelis, D. & Travis, C. Endemic disease in environments with spatially heterogeneous host populations. Math. Biosci. 63, 289–302 (1983).
MathSciNet  MATH  Google Scholar 
May, R. M. & Anderson, R. M. Spatial heterogeneity and the design of immunization programs. Math. Biosci. 72, 83–111 (1984).
MathSciNet  MATH  Google Scholar 
Hethcote, H. W. & Van Ark, J. W. Epidemiological models for heterogeneous populations: proportionate mixing, parameter estimation, and immunization programs. Math. Biosci. 84, 85–118 (1987).
MathSciNet  MATH  Google Scholar 
Lloyd, A. L. & May, R. M. Spatial heterogeneity in epidemic models. J. Theor. Biol. 179, 1–11 (1996).
ADS  CAS  PubMed  Google Scholar 
SafeGraph. SafeGraph Social Distancing Metrics. https://docs.safegraph.com/docs/social-distancing-metrics (2021).
Dong, E., Du, H. & Gardner, L. An interactive web-based dashboard to track COVID-19 in real time. Lancet Infect. Dis. 20, 533–534 (2020).
CAS  PubMed  PubMed Central  Google Scholar 
Mehta, M. Random Matrices, 3rd ed, Sec. 1.5 (Elsevier, 2004).
Akemann, G., Baik, J. & Di Francesco, P. The Oxford Handbook of Random Matrix Theory (Oxford University Press, 2011).
MATH  Google Scholar 
Livan, G., Novaes, M. & Vivo, P. Introduction to Random Matrices: Theory and Practice (Springer, 2020).
MATH  Google Scholar 
Rosenzweig, N. & Porter, C. Repulsion of energy levels in complex atomic spectra. Phys. Rev. 120, 1698–1714 (1960).
ADS  CAS  Google Scholar 
Girvan, M. & Newman, M. E. Community structure in social and biological networks. Proc. Natl. Acad. Sci. 99, 7821–7826 (2002).
ADS  MathSciNet  CAS  PubMed  PubMed Central  MATH  Google Scholar 
Luo, F., Zhong, J., Yang, Y., Scheuermann, R. H. & Zhou, J. Application of random matrix theory to biological networks. Phys. Lett. A 357, 420–423 (2006).
ADS  CAS  Google Scholar 
DeMeo, P., Ferrara, E., Fiumara, G. & Ricciardello, A. A novel measure of edge centrality in social networks. J. Knowl.-Based Syst. 30, 136–150 (2012).
Google Scholar 
Brody, T. A. et al. Random-matrix physics: Spectrum and strength fluctuations. Rev. Mod. Phys. 53(385–479), 391 (1981).
ADS  MathSciNet  Google Scholar 
Guhr, T., Müeller-Groeling, A. & Weidenmüller, H. A. Random matrix theories in quantum physics: Common concepts. Phys. Rep. 299(189–425), 228 (1998).
MathSciNet  Google Scholar 
Hakke, F. Quantum Signatures of Chaos 3rd edn, Sec 4.5 (Springer, 2010).
Naether, U., Postnikov, E. & Sokolov, I. Infection fronts in contact disease spread. Eur. Phys. J. B 65, 353–359 (2008).
ADS  CAS  Google Scholar 
Dee, G. & Langer, J. Propagating pattern selection. Phys. Rev. Lett. 50, 383 (1983).
ADS  Google Scholar 
Ben-Jacob, E., Brand, H., Dee, G., Kramer, L. & Langer, J. Pattern propagation in nonlinear dissipative systems. Phys. D: Nonlinear Phenomena 14, 348–364 (1985).
ADS  MathSciNet  MATH  Google Scholar 
Bramson, M. Convergence of Solutions of the Kolmogorov Equation to Travelling Waves, Vol. 285 (American Mathematical Soc., 1983).
MATH  Google Scholar 
Paquette, G., Chen, L.-Y., Goldenfeld, N. & Oono, Y. Structural stability and renormalization group for propagating fronts. Phys. Rev. Lett. 72, 76 (1994).
ADS  CAS  PubMed  Google Scholar 
Ebert, U. & van Saarloos, W. Front propagation into unstable states: Universal algebraic convergence towards uniformly translating pulled fronts. Phys. D Nonlinear Phenomena 146, 1–99 (2000).
ADS  MathSciNet  MATH  Google Scholar 
Birzu, G., Hallatschek, O. & Korolev, K. S. Fluctuations uncover a distinct class of traveling waves. Proc. Natl. Acad. Sci. 115, E3645–E3654 (2018).
ADS  MathSciNet  CAS  PubMed  PubMed Central  MATH  Google Scholar 
Sen, P., Yamana, T. K., Kandula, S., Galanti, M. & Shaman, J. Burden and characteristics of COVID-19 in the united states during 2020. Nature 598, 338–341 (2021).
ADS  CAS  Google Scholar 
SafeGraph. SafeGraph Neighborhood Patterns. https://docs.safegraph.com/docs/neighborhood-patterns (2021).
SafeGraph. SafeGraph Open Census Data. https://docs.safegraph.com/docs/open-census-data (2021).
Download references
The authors acknowledge the generous support of the Chan Zuckerberg Initiative and Chan Zuckerberg Biohub. The authors also thank Mark Rychnovsky for useful conversations on random matrix theory in the early stages of this project, and Naomi Rankin for her help with a preliminary version of the lattice simulations. D.Y. acknowledges support by MINECO (Spain) through Grant No. PGC2018-094684-B-C21, partially funded by the European Regional Development Fund (FEDER). We also are grateful to the anonymous reviewers of Scientific reports who asked deep questions, that helped us to better understand our results.
Chan Zuckerberg Biohub, 499 Illinois Street, San Francisco, CA, 94158, USA
Guillaume Le Treut, Greg Huber, Mason Kamb, Aaron McGeever & David Yllanes
Department of Physics, Kadanoff Center for Theoretical Physics, University of Chicago, Chicago, IL, 60637, USA
Kyle Kawagoe
Okinawa Institute of Science and Technology, Onna-son, Okinawa, 904-0495, Japan
Jonathan Miller & Reuven Pnini
Chan Zuckerberg Initiative, Redwood City, CA, 94063, USA
Boris Veytsman
School of Systems Biology, George Mason University, Fairfax, VA, 22030, USA
Boris Veytsman
Instituto de Biocomputación y Física de Sistemas Complejos (BIFI), 50018, Zaragoza, Spain
David Yllanes
You can also search for this author in PubMed Google Scholar
You can also search for this author in PubMed Google Scholar
You can also search for this author in PubMed Google Scholar
You can also search for this author in PubMed Google Scholar
You can also search for this author in PubMed Google Scholar
You can also search for this author in PubMed Google Scholar
You can also search for this author in PubMed Google Scholar
You can also search for this author in PubMed Google Scholar
You can also search for this author in PubMed Google Scholar
The concept for this research grew out of discussions among all authors; G.L.T. performed research; G.H., A.M., M.K. analyzed data; K.K., J.M., R.P., performed analytical/numerical computations; G.L.T., B.V. and D.Y. wrote the paper, with contributions from all authors.
Correspondence to Guillaume Le Treut or Greg Huber.
The authors declare no competing interests.
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Supplementary Video S1.
Supplementary Video S2.
Supplementary Video S3.
Supplementary Video S4.
Supplementary Video S5.
Supplementary Video S6.
Supplementary Video S7.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.
Reprints and Permissions
Le Treut, G., Huber, G., Kamb, M. et al. A high-resolution flux-matrix model describes the spread of diseases in a spatial network and the effect of mitigation strategies. Sci Rep 12, 15946 (2022). https://doi.org/10.1038/s41598-022-19931-w
Download citation
Received:
Accepted:
Published:
DOI: https://doi.org/10.1038/s41598-022-19931-w
Anyone you share the following link with will be able to read this content:
Sorry, a shareable link is not currently available for this article.

Provided by the Springer Nature SharedIt content-sharing initiative
By submitting a comment you agree to abide by our Terms and Community Guidelines. If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.
Advertisement
Advanced search
Scientific Reports (Sci Rep) ISSN 2045-2322 (online)
© 2022 Springer Nature Limited
Sign up for the Nature Briefing newsletter — what matters in science, free to your inbox daily.

source

Download Point Blank’s Free Vocoder Plugin – Watch Tutorial Now – Magnetic Magazine

Homer Dudley initially developed vocoders at the Bell Labs in the 1930s, whilst he was experimenting with speech synthesis. Vocoders work using a modulator signal, like a voice or drum beat, and a carrier signal, such as a sustained pad or synth sound. 
These signals are then split through a number of bandpass filters that combine and output the common frequency between the two signals. An iconic use of that robotic-vocal sound, created by a vocoder, can be found in Kraftwerk’s 1978 track, ‘The Robots‘.
If you’re keen to learn more music production skills, much like this one, then be sure to check out Point Blank’s music production courses in Los Angeles. Alternatively, if you’re looking to study with Point Blank in London or online, they have a range of degree courses that specialize in music production and sound engineering. Check out their full roster here.
During the video, Dan demonstrates how PB’s new vocoder works, showing how you can transform a drum beat into a robotic vocal sound. He runs through the different parameters, revealing how you can get the most out of Point Blank’s free vocoder plugin.
The new Max for Live plugin by Dan Herbert is available exclusively from Point Blank – it will only cost you your email address. To find help on how to install the new Max for Live vocoder plugin, be sure to watch their previous Kick Drum Designer video.

source

Researchers Uncover Years-Long Mobile Spyware Campaign Targeting Uyghurs – The Hacker News

A new wave of a mobile surveillance campaign has been observed targeting the Uyghur community as part of a long-standing spyware operation active since at least 2015, cybersecurity researchers disclosed Thursday.
The intrusions, originally attributed to a threat actor named Scarlet Mimic back in January 2016, is said to have encompassed 20 different variants of the Android malware, which were disguised as books, pictures, and an audio version of the Quran.
The malware, while relatively unsophisticated from a technical standpoint, comes with extensive capabilities to steal sensitive data from an infected device, send SMS messages on the victim’s behalf, make phone calls, and track their locations.
Additionally, it allows the recording of incoming and outgoing phone calls as well as surrounding audio.
“All this makes it a powerful and dangerous surveillance tool,” Israeli cybersecurity firm Check Point said in a technical deepdive, calling the spyware MobileOrder.
It’s worth noting that a part of the campaign was recently disclosed by researchers from the MalwareHunterTeam and Cyble, in which a book written by the exiled Uyghur leader Dolkun Isa was used as a lure to deliver the malware.
Check Point said it observed MobileOrder artifacts in the wild right from 2015 to mid-August 2022, with the exception of 2021, when none were detected.
Attack campaigns likely involve the use of social engineering tactics to trick unsuspecting victims into launching malicious applications that reference seemingly innocuous documents, photos, and audio files.
These apps contain a variety of baits, including a PDF about guerrilla warfare and pictures related to the deployment of paramilitary forces in Ürümqi, the capital of the Xinjiang Uyghur Autonomous Region, in the aftermath of the deadly April 2014 attack.
Opening the rogue app, in turn, launches a decoy document designed to distract the target from noticing the malicious actions in the background.
“Some of the versions also ask for Device Admin and root access, which not only gives the malware full access to the device, but also prevents the victim from easily uninstalling the application,” the researchers said.
Other features supported by MobileOrder include executing a remote shell and even dropping additional Android Package (APK) files.
The campaign’s attribution to Scarlet Mimic, per Check Point, stems from clear code overlaps, shared infrastructure, and the same victimology patterns.
Furthermore, the ongoing use of MobileOrder signals a shift in attack vector from desktop to mobile surveillance, what with the actor previously linked to a Windows malware called Psylo Trojan.
While it’s not clear which of these attacks throughout the past seven years have been successful, the very fact that the malware authors are continuing to deploy the spyware is an indication that some of these efforts have paid off.
“The persistence of the campaign, the evolution of the malware, and the persistent focus on targeting specific populations indicate that the group’s operations over the years are successful to some extent,” Check Point said.
Sign up for cybersecurity newsletter and get latest news updates delivered straight to your inbox daily.

source

Judge Names Special Master and Keeps Block on Mar-a-Lago Files Inquiry – The New York Times

Advertisement
Supported by
The Justice Department is planning to appeal, but the decision is likely to significantly delay its investigation into former President Donald J. Trump’s handling of government records.
Send any friend a story
As a subscriber, you have 10 gift articles to give each month. Anyone can read what you share.
Charlie Savage, Alan Feuer and
WASHINGTON — A federal judge on Thursday rejected the Justice Department’s request to resume a key part of its inquiry into former President Donald J. Trump’s handling of sensitive government records and appointed an outside arbiter to review thousands of documents seized last month from his Florida residence.
The judge, Aileen M. Cannon, declined to lift any part of an order she issued last week that barred the department from using the documents, including about 100 marked classified, in its investigation until the arbiter, known as a special master, had completed a review.
In her 10-page decision, Judge Cannon appointed a special master suggested by the Trump legal team and agreed upon by the government: Raymond J. Dearie, a semiretired judge from the Federal District Court for the Eastern District of New York.
Judge Dearie will now have the authority to sift through more than 11,000 records the F.B.I. carted away from Mr. Trump’s estate, Mar-a-Lago, on Aug. 8. The move was a blow to the Justice Department, almost certain to significantly delay its investigation into whether the former president unlawfully retained national defense records or obstructed repeated attempts by federal officials to retrieve them.
The department had asked Judge Cannon to lift restrictions on its use of documents with classification markings and set a Thursday deadline for her to respond before it said it would ask an appeals court to intervene. The department is now planning to appeal the decision, and top officials were meeting to discuss the timing of their filing, according to a senior law enforcement official.
Judge Cannon’s order was the latest turn in what has now become a protracted and politically fraught court battle over the trove of documents seized in the search of Mar-a-Lago.
While Judge Cannon’s order was a victory for Mr. Trump and his legal team, she also made two significant concessions to the government.
She said, for instance, that Judge Dearie should first look at the classified documents and “thereafter consider prompt adjustments to the court’s orders as necessary.” That raised the possibility that Judge Dearie might quickly clear the items and the F.B.I. would then regain unrestricted use of them in its criminal inquiry. In her order, Judge Cannon said the special master should try to finish his review by Nov. 30.
Judge Cannon also clarified her temporary ban on the Justice Department’s investigation. While the department could not present the seized documents to a grand jury or ask witnesses about their contents, she said, the F.B.I. could continue to investigate “the movement and storage of seized materials, including documents marked as classified, without discussion of their contents.”
Still, Judge Cannon, whom Mr. Trump appointed in November 2020, declined to back down on having the special master vet those documents, too — and making the F.B.I. wait, at least for now, to have full access to them in its investigation.
In her order, Judge Cannon took pains to describe herself as mindful of the traditional reluctance of the judicial branch to intrude on executive authority in national security matters. But she wrote that her court could not “abdicate its control over questions of privilege and does not find the government’s argument sufficiently convincing as presented.”
Among other things, she went on, the government had not suggested there was any “identifiable emergency or imminent disclosure of classified information arising from” Mr. Trump’s retention of the materials. Instead, she added, “the unwarranted disclosures that float in the background have been leaks to the media after the underlying seizure.”
In naming Judge Dearie as special master, Judge Cannon did not address an issue the Justice Department had flagged in saying it was open to his appointment: As a semiretired judge on senior active status, he occasionally hears cases, and active judges are not supposed to have outside employment.
Judge Cannon said Mr. Trump would pay the entire cost of Judge Dearie’s compensation and expenses, along with the cost of support staff or advisers he needs, given that Mr. Trump had requested a special master.
Asked about the possibility of being indicted in the investigation, Mr. Trump told a conservative radio host on Thursday that there would be “problems in this country the likes of which perhaps we’ve never seen before.”
Last week, Judge Cannon issued an order announcing that she would appoint a special master to go over the materials to set aside any potentially covered by attorney-client privilege or, in a legally disputed move, executive privilege. She also prevented prosecutors from using any of the records in its investigation in the meantime.
But as part of her order, Judge Cannon permitted the government to continue using the documents for a national security risk assessment and a classification review.
Days later, the Justice Department asked Judge Cannon to stay parts of her order, requesting that the F.B.I. be allowed to resume investigating matters related to about 100 of the seized files marked as classified. The department argued that the bureau’s criminal investigation of the files was “inextricably linked” to the separate national-security reviews that she had said could continue.
At the same time, the Justice Department told Judge Cannon that if she did not agree to its requests for a stay by Thursday, it would ask the appeals court to intervene.
In effect, the government had sought a compromise, acquiescing to Judge Cannon’s decision to impose a special master and, at least temporarily, to bar F.B.I. agents and prosecutors from working further with the overwhelming majority of the seized documents and other records. In return, the Justice Department asked Judge Cannon to let its investigators continue working with a smaller batch.
To make their point, lawyers for the Justice Department singled out nearly 50 empty folders bearing classified banners that were found during the search of Mar-a-Lago. The lawyers said the F.B.I. needed to investigate what happened to the contents of the folders so intelligence officials could determine whether national security had been compromised.
The Office of the Director of National Intelligence has also said that after Judge Cannon’s order, it paused its risk assessment and classification review. In her ruling on Thursday, she expressed doubt about whether the government could no longer conduct such an assessment without the criminal investigation proceeding — even as she put forward a very narrow interpretation of her initial order.
Judge Cannon also responded to criticism that she had put too much weight on Mr. Trump’s status as a former president in deciding that a special master with a broad mandate was justified. Critics have said that she violated the principle that the law is supposed to be evenhanded and treat everyone equally and that she has shown him special deference.
She wrote that “the principles of equity require the court to consider the specific context at issue, and that consideration is inherently impacted by the position formerly held” by Mr. Trump.
She added that she steadfastly believed that appointing a special master to review the seized materials and temporarily barring the government from using them was principled and necessary “to ensure at least the appearance of fairness and integrity under unprecedented circumstances.”
Advertisement

source

Russia-Ukraine war latest updates: Kremlin stages referendum vote – The Washington Post

Sign in
Sign in
Hundreds of people were arrested Saturday during demonstrations in Russia against the nation’s “partial mobilization” of troops that the Kremlin plans to throw into its invasion of Ukraine. Some who had participated in demonstrations earlier in the week had been given military summonses, state media reported.
The Kremlin’s staged referendums in Ukraine — which some residents called a vote “under a gun barrel” — were underway for a second day and were scheduled to continue until Tuesday in regions controlled by Russian forces. Western leaders have condemned the orchestrated referendums as a pretext to annex swaths of the country.
Here’s the latest on the war and its ripple effects across the globe.
Propaganda newspapers show how Russia promoted annexation in Kharkiv: Over the months Russian troops occupied Izyum in Ukraine’s northeast, puppet authorities regularly distributed propaganda newspapers to residents, pushing a narrative of normalcy and unity even as homes and infrastructure were demolished, stores were looted, and civilians struggled to find basic provisions to survive.
A trove of the Russian-language newspapers, provided to The Washington Post by a resident who said he kept them “for history,” paints a surreal version of events on the ground running in near total contradiction to the narrative both from the Ukrainian government in Kyiv and accounts from residents who survived the violent takeover of the city in March.
Ukrainian forces recaptured Izyum in a surprise counteroffensive earlier this month, Siobhán O’Grady and Sergii Mukaieliants report, sparing the city from a staged referendum in an attempt to justify Russian annexation like those that began Friday in parts of Luhansk, Donetsk, Kherson and Zaporizhzhia regions.
The propaganda newspapers show how Russian forces tried to take advantage of the city’s information vacuum during the occupation when cellphone and internet service was mostly cut. The papers sought to evoke nostalgia among civilians for the Soviet era, to turn residents against Ukrainian forces and to promote deep historical and cultural ties with Russia, apparently in preparation for annexation.
Mary Ilyushina contributed to this report.

source

How to Edit a Podcast in GarageBand

Liv Hanna releases a delightful neo-soul single, entitled, “Boy” – Bong Mines Entertainment

Liv Hanna is an up-and-coming singer-songwriter from Long Island, New York. Not too long ago, she released a delightful neo-soul single, titled, “Boy.”

“My inspiration for the song ‘Boy’ came about from observing bad behavior in my peer group and how it relates to self-respect. ‘Boy’ is about not allowing a toxic relationship to dictate you or even ruin your life. I tried to convey a message which implies whoever is listening is better off without their significant other. It encourages independence and standing up for oneself.” – Liv Hanna explained
‘Boy’ contains a relationship-based narrative and an easily absorbing vibe that will surely capture listeners’ attention. The cushy tune possesses a driving bassline groove and layered vocal harmonies capped by Liv Hanna’s sultry delivery and insightful lyrics. Written by Hanna with her co-producers, Guy Eckstine and Paul Pesco, the music production draws listeners into Hanna’s sonic realm with catchy hooks and cerebral sonic atmospheres. Furthermore, “Boy” serves as a tasty appetizer to what listeners can expect to hear from Liv Hanna as she continues to release new music and videos throughout 2022.
“I use my music to convey experiences that will make my audience feel something real. I connect music to everything that I go through, and I know from first-hand experience how much a song can influence a mood and help someone. I think that’s why I’ve always been so drawn to music because no matter what is happening, I’m always pushed to feel something new when I listen to a song.” – Liv Hanna explained
When the Covid-19 pandemic took hold, Liv Hanna found herself isolated from her classmates while sequestered at home. During that time, she began writing songs and released “New You,” produced by Guy Eckstine and Joe “Legacy” Lindsay. As Hanna continued to compose new songs, Eckstine called on session guitar ace, Paul Pesco (Madonna, Jennifer Lopez) to co-produce. The collaboration led to the release of the singles “Bedroom,” and “Self Made,” both showcasing Hanna’s evolution towards incorporating R&B influences into her music.
Liv Hanna - “Boy” cover art

We recommend adding Liv Hanna’s “Boy” single to your favorite neo-soul playlist. Also, let us know how you feel in the comment section below. Mel Blanc’s famous catchphrase, That’s All Folks! Thanks for reading another great article on Bong Mines Entertainment – the hot spot for new music. Always remember that (P) Positive, (E) Energy, (A) Always, (C) Creates, (E) Elevation (P.E.A.C.E). Contact us on social media via Twitter, YouTube, Instagram, or Facebook, and let us know what’s going on. Disclaimer: This post, contains affiliate links, and we may be compensated if you make a purchase through links on our website.

Your email address will not be published.







Forgotten Password

source

New roles in virtual production and how to get them – NewscastStudio

‘Tamron Hall’ gets new logo, other updates for Season 4
‘Jennifer Hudson’ retools ‘Ellen’ set pieces, creating inconsistent look
‘Sherri’ overhauls ‘Wendy Williams’ space with distinct look
Category Sponsor
Like sands through the hourglass, ‘NBC News Daily’ continues ‘Day’-driven motif
U.S. morning newscasts continue extended coverage of Queen Elizabeth II
Anchor’s tweet provides clues about ‘NBC News Daily’ set, graphics
Category Sponsor
CNN climate segment falls flat
Viewers get top deck tour of new Chinese aircraft carrier thanks to AR
Asharq outlines explosive Ukrainian land mine dangers using AR segment
Category Sponsor
Jib captures fairytale-like closing shot of Johnson’s rocky tenure as PM
Viral video shows couple’s dances choreographed to NFL theme music
Amazon reveals ‘Thursday Night Football’ theme music
Category Sponsor
CBS Sports goes cinematic in new hallmark
‘SportsCenter’ adds 38 million pixels in Studio X upgrade
ESPN on ABC debuting new look, sound hallmarks during Little League World Series
Category Sponsor
Industry Insights: Studio robotics and camera control, part two
Industry Insights: Studio robotics and camera control, part one
Sony product updates focused on imaging, IP, cloud and visualization workflows
Category Sponsor
Illinois station remains on air even as its studio floods
CBS News Detroit prepares for the elements with Next Weather Tracker
Vizrt Group acquires Flowics HTML5 graphics platform
CBS News Detroit prepares for the elements with Next Weather Tracker
Nexstar’s WKRG 5 adds mobile weather vehicle, ‘The Beast’
Advisors Excel adds corporate broadcast studio
Al Jazeera’s Studio 5 features nearly 200 Ayrton Diablo lights
Case Study: How Bill Holshevnikoff lit Telemundo Sacramento
Litepanels launches 2×1 Gemini panel light with 23,000 lux of output
Category Sponsor
Vizrt Group acquires Flowics HTML5 graphics platform
Top reasons an integrator should build your broadcast facility furniture
Column: The status of SDI to IP transitions and the benefits of maintaining hybrid environments… for now
Category Sponsor
Subscribe to NewscastStudio’s newsletter for the latest in broadcast design and engineering delivered to your inbox.
From the types of jobs on offer, to what skills and experience you need to join them, here’s your essential primer in joining the rapidly evolving virtual production industry.
Virtual production (VP) is rapidly becoming a popular approach across film and TV, live events, and immersive media experiences. Not only are there practical benefits – saving both time and cost when compared to expensive location shoots – but it also opens up a whole new world of creative possibilities for any kind of media production.
With the industry rapidly evolving, roles in VP are becoming increasingly in demand, opening up a range of opportunities for both newcomers and established industry professionals. In this article, we’re sharing some examples of key roles in VP and discussing tips on breaking into the industry as a newcomer.

With these new roles constantly evolving, it’s clear that there is plenty of opportunity to get involved with the ever-expanding virtual production world. But how do you actually get started with a career in virtual production?
As with any industry, networking is a good place to start. However, it’s also a good idea to familiarize yourself with the virtual production pipeline. Having an understanding of each element of the production, even those you’re not directly involved with – such as video I/O, real-time animation, technical support, and stage skills – will give you an advantage in creating a seamless workflow. And if you work on the creative side, then a showreel is essential for showcasing your work.
Teamwork and communication are particularly important and often undervalued skills for the virtual production environment, says Erik Beaumont, head of mixed reality (MR) at The Famous Group (TFG), a production company specializing in virtual and augmented reality for sports and live events. 

Erik also highlights the advantage of having a traditional production background. “Production timelines and deadlines are stacked slightly differently. We emphasize different things like how the dailies work and cutoff dates for approvals, but if you have the basic toolset, you’ll adjust fairly seamlessly,” says Erik.
He also advises that knowing your way around a real-time engine will put you in a good position, and that this is not as hard as it sounds. “Pick a game engine of choice and just do something in it. People imagine a high threshold of how complicated it is, but the reality is different,” says Erik.
Spanish 3D artist Fernando Trueba Llano, a technical artist at Televisión Canaria, agrees that real-time engines are a good starting point for those who want to get into virtual production but don’t know where to start. 
“My recommendation would be to start studying the basics of 3D and focus on Unreal Engine,” says Fernando. “Dedicate personal time to dig in and learn as much as possible. […] And finally, put into practice what you have learned and create demos of your capabilities with the engine.”

If you’re interested in gaining new skills in VP from a traditional production background, or if you’d like to transfer your experience in gaming, VFX, or real-time engines, here’s a taster of some of the new jobs and skills currently needed in the ever-evolving virtual production industry.
The Virtual Production Supervisor is a leadership role, taking charge of the entire virtual production set. People in this role can expect to act as a bridge between different sections of the crew. This requires strong people skills – and you’ll also need an excellent understanding of the tech involved. This is an ideal role for producers, or those with a background in VFX, such as a VFX supervisor.
A virtual production volume – otherwise known as an XR volume or VP stage – is made up of many large LED panels that display media content. This can be anything from stills and 2D footage to real-time 3D graphics generated by a game engine.
These volumes need someone to oversee the content piped into these screens, where the Volume Operator comes in. Often working with assistants, depending on the production size, the Volume Operator is usually someone with experience in creating real-time graphics on gaming engines such as Unreal Engine or Unity.
As well as having someone to operate the volume, virtual productions require someone to manage and maintain the screens themselves. LED panels are a key piece of the rig in a VP environment so it is essential to have an LED Engineer, otherwise known as an LED Technician, on-set to ensure they work correctly and consistently.
LEDs have been used across live events and broadcasts for some time, so many AV technicians will find they already have the required skills to become an LED Engineer. Other relevant skills include experience in motion capture and camera tracking systems, plus knowledge of Unreal Engine and other real-time tools.
In the visual effects industry, VFX supervisors take charge of an entire project, managing the people, workflows, and timelines for delivery. This is not an entirely new role, but has evolved within the virtual production world. 
In VP, a VFX Supervisor is responsible for all the visual effects on a real-time production. This role works across the creative side but also has a solid understanding of the technical aspects involved, especially newer innovations like real-time technology. 
Another leadership role, this one focuses on motion capture. Like the VFX Supervisor, this role has existed for some time but has now expanded to incorporate the real-time environment. Again, working knowledge of motion capture and the ability to work with camera tracking systems is essential.   
As the name suggests, this person operates a virtual camera using an external device such as a tablet. A good understanding of cinematography is important, as well as the ability to keep up to date with all the latest equipment and software on offer. This is also an excellent role for people with experience in creating computer-generated animations using game engines. 
Real-time game engines, such as Unreal Engine and Unity, are becoming an increasingly essential part of the virtual production pipeline. The Engine Operator is responsible for operating the real-time engine being used in the production.
About
Carina Schoo, Pixotope
Carina is an experienced communications professional with a track record in the marketing of immersive imaging technologies and virtual solutions. Carina joined the virtual production software company Pixotope as Education Program Manager. Her primary responsibility is to design, manage and execute the Pixotope Education Program (PEP). As a mediator between educational institutions and the virtual production industry, she identifies shortages and needs on both sides, breaking down barriers to the wide adoption and growth of virtual production. Before joining Pixotope, Carina worked as a Project Manager for one of the renowned German Fraunhofer Institutes, the Fraunhofer Heinrich Hertz Institute in Berlin (DE), and as Communications Consultant for VRINN Immersive Learning Cluster, Hamar (NO). Carina holds a Master’s degree in Communications and Languages from the Technische Universität Berlin.
, , , , , ,
Broadcast Automation, Broadcast Engineering News, Featured, Featured Column, IP Based Production, Jobs, Virtual Production 101

Copyright © 2003-2022 NewscastStudio, an HD Media Ventures LLC company. All rights reserved. NewscastStudio, The trade publication broadcast production and related marks and trade dress are marks of NewscastStudio.
Never see this message again.

source

Optimal selective floor cleaning using deep learning algorithms and reconfigurable robot hTetro | Scientific Reports – Nature.com

Thank you for visiting nature.com. You are using a browser version with limited support for CSS. To obtain the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles and JavaScript.
Advertisement
Scientific Reports volume 12, Article number: 15938 (2022)
Metrics details
Floor cleaning robots are widely used in public places like food courts, hospitals, and malls to perform frequent cleaning tasks. However, frequent cleaning tasks adversely impact the robot’s performance and utilize more cleaning accessories (such as brush, scrubber, and mopping pad). This work proposes a novel selective area cleaning/spot cleaning framework for indoor floor cleaning robots using RGB-D vision sensor-based Closed Circuit Television (CCTV) network, deep learning algorithms, and an optimal complete waypoints path planning method. In this scheme, the robot will clean only dirty areas instead of the whole region. The selective area cleaning/spot cleaning region is identified based on the combination of two strategies: tracing the human traffic patterns and detecting stains and trash on the floor. Here, a deep Simple Online and Real-time Tracking (SORT) human tracking algorithm was used to trace the high human traffic region and Single Shot Detector (SSD) MobileNet object detection framework for detecting the dirty region. Further, optimal shortest waypoint coverage path planning using evolutionary-based optimization was incorporated to traverse the robot efficiently to the designated selective area cleaning/spot cleaning regions. The experimental results show that the SSD MobileNet algorithm scored 90% accuracy for stain and trash detection on the floor. Further, compared to conventional methods, the evolutionary-based optimization path planning scheme reduces 15% percent of navigation time and 10% percent of energy consumption.
Generally, human traffic is heavy on commercial premises like shopping malls, food courts, community clubs, and hospitals. Therefore, the chances of accumulating trash, stains (footprint stains and other forms of stains), and debris are quite high, raising safety and hygiene concerns. Hence, frequent cleaning is essential in commercial premises in pandemic situations like COVID-19. Due to long working hours, low wages, and unwillingness to work as a cleaner, workforce shortage has been a constant problem for cleaning and maintenance tasks in commercial premises1,2.
In recent years, mobile cleaning robots have been widely used to clean and maintain commercial premises and help overcome the workforce shortage. It performs the cleaning task in public premises using two cleaning modes: auto cleaning/scheduled cleaning mode and selective area cleaning/spot cleaning mode2,3,4,5. Here, the auto cleaning/scheduled cleaning mode cleans the given workspace autonomously once a day or at a scheduled time interval using a pre-build floor map. In selective area cleaning/spot cleaning mode, the robot will clean only dirty areas instead of cleaning the whole region. In contrast with auto cleaning/scheduled cleaning mode, the selective area cleaning/spot cleaning mode could save energy by avoiding non-dirty areas, reducing the cleaning time, and enhancing the lifetime of the robot6. However, the selective area cleaning/spot cleaning mode hasn’t been designed with autonomous operation, where the robot requires human assistance to perform the selective area cleaning/spot cleaning. Moreover, human-assisted selective area cleaning or spot cleaning is labor-intensive and time-consuming for larger workspaces. The cleaning supervisor needs to continuously monitor or inspect the environment to enable selective area cleaning/spot cleaning mode (according to human traffic in the environment or trash and stains detection).
Computer vision-based autonomous selective area cleaning/spot cleaning is a viable solution for overcoming the shortcomings mentioned above. Recently, Machine Learning (ML) and Deep Learning (DL) based computer vision algorithms are widely used for mobile cleaning robots for stain and trash detection and path planning tasks7,8,9. However, selective area cleaning/spot cleaning using onboard robot vision systems and generating the optimal path generation has a lot of practical challenges. Searching and finding the dirty region itself increases the energy consumption and processing time by scanning the entire area of operation. Further, monitoring human traffic from a robot perspective and computing the selective area cleaning/spot cleaning region took more computation time and power.
Hence, there is a need for a novel and scalable approach to overcome the challenges mentioned above in the selective area cleaning/spot cleaning method. Therefore, this article proposes a novel selective area cleaning/spot cleaning framework for indoor floor cleaning robots using RGB-D vision sensor based Closed Circuit Television (CCTV) network, deep learning algorithm, and optimal complete waypoints path planning algorithm. Furthermore, the main contributions of this paper are divided threefold:
Design and development of selective area cleaning/spot cleaning using deep learning-based computer vision algorithm and RGB-D vision sensor network. The deep learning algorithms find the selective area cleaning/spot cleaning region based on detecting stains and trash on the floor and tracing the human traffic patterns (common space: hall, entrance, and hallway/entryway).
Compute the 3D location of stain and trash in world coordinates to estimate the complete waypoint coverage path planning by evolutionary methods to drive the cleaning robot to a selective area cleaning/spot cleaning.
Validate the efficiency of selective area cleaning/spot cleaning framework in the real workspace with in-house developed reconfigurable floor cleaning robot hTetro.
The manuscript is organized as follows—section “Related work” elaborates the related works, and the proposed system is presented in section “Overview of the proposed system”. Finally, section “Experimental results” provides the experimental results, followed by the conclusion and remarks of this work.
This section presents the various dirt, stain, and trash detection schemes reported in the literature with detail summery (Table 1) and the optimal path planning methods used for cleaning robots. In10 Andersen et al. proposed a floor cleaning robot with a cleanness sensor using a computer vision algorithm and LED illumination. The sensor identifies dirty areas by analyzing surface images and producing pixel-by-pixel cleanness maps, applying a multivariable statistical approach. Borman et al.11,12 proposed a spectral residual filtering-based method for dirt detection. The dust threshold for various floor patterns is fixed based on maximal and minimal filter response. However, the dirt detection ratio is varied upon surface types. Mud and dirt on the floor are identified and classified for cleaning robots using Maximally Stable Extremal Regions (MSER) and Minimal Canny edge detection algorithms13. Lee and Banerjee14 suggested a solution for the overall cleaning schedule, which also generates an effective path for each cleaning cycle. Simulation-based Optimisation (SO) technique is applied for floor cleaning to avoid redundant paths and maintain tolerable dirt particle levels. In6 vision-based dirt detection was proposed for floor cleaning robot hTetro. Here, the authors use a periodic pattern detection filter to identify the dirt region on the floor and a tilling algorithm for generating the cleaning pattern. Grunauer et al.15 proposed an algorithm to detect dust spots based on Gaussian Mixture Model (GMM) applied to RGB images. This supports distinguishing the dust spots from the floor surface. Martinez et al.16 proposed a vision-based dust detection technique for the planar surface. The median filter and the thresholding technique were applied to distinguish the dust and the floor surface. Further, RANSAC algorithm was adopted to detect plane surfaces. However, as mentioned above, the scheme has significant practical limitations; for instance, the heavy detection of dirt depends on the floor background and cannot distinguish the dirt classes like trash or stain11,12,13.
Deep Learning (DL) based object detection is an emerging technique. Nowadays, it has been widely utilized in the autonomous robot vision system, abandoned object detection in surveillance applications8,9, human and vehicle tracking applications17,18, to detect stains and trash in cleaning robots, classify garbage in waste management. In19 author evaluate the efficiency of various deep learning-based object detection framework for marine trash detection. The authors report that the Single Shot Detector (SSD) architecture has high detection accuracy than the Yolo v2 and Tiny Yolo object detection models. GoogLeNet-Overfeat object detection model was trained for outdoor trash detection proposed by Rad et al.20. Here, the detection framework was trained with 18000 trash images containing waste papers, metal/plastic cans, and leaves. The trained model obtained a maximum of 77.3 % precision for detecting trash on the street. Ramalingam et al.7 proposed a debris detection and classification framework for floor cleaning robot application. The authors7 train the SSD MobileNet object detection algorithm with debris image dataset to detect and classify the trash and stains on the floor. Zhihong et al.,21 proposed an automatic trash sorting system using Region Proposal Network (RPN) and VGG-16 deep learning model. The authors report that the trained model had 3% miss detection and 9% false detection. Gaurav et al.21, proposed a smartphone app-based outdoor garbage localization technique where AlexNet CNN was trained to detect the outdoor garbage and detect the outdoor trash with (87%) detection accuracy. Yin et al.2 proposed Human Support Robot (HSR) based on deep learning for cleaning the tables. A darknet framework-based 16-layer CNN framework was developed. The model achieved 95% accuracy in detecting food strains and trash on the surface of the table. Further, a similar autonomous robot was developed using deep learning architecture, which has the capability to produce two-dimension trajectories from human kinesthetic instances22,23.
Apart from dirt, stain, and trash detection, path planning is also a key function for floor cleaning robots to achieve efficient cleaning. In the literature, path planning for cleaning robots was widely studied. Generally, spiral and zigzag patterns are widely used to generate the path for cleaning robots, aiming to maximize the area’s coverage. However, these patterns have significant limitations in covering the constrained areas, such as narrow space, and are only suitable for fixed morphology platforms7. Further, computing the optimal shortest path is another key challenge for the selective area cleaning/spot cleaning task. Reconfigurable morphology robots can cover the maximum area and arbitrary workspace compared with a fixed morphology robot and are more suitable for selective/ spot-cleaning tasks. Self-reconfigurable robots have attracted significant attention, and the development of autonomy systems mostly focused on autonomous motion control and mechanism design. There are only a few works in path planning for self-reconfigurable robots, e.g., the work in24 proposes a complete coverage path planning model trained using deep reinforcement learning (RL) for the tetromino based reconfigurable robot platform to simultaneously generate the optimal set of shapes for any pretrained. Research on path planning so far focuses on fixed-morphology robots and considered robots as a single point. But, representative path planning algorithms can be broadly classified as classical graph search algorithms (Dijkstra, A*, D*, and D* Lite)25,26,27,28, bio-inspired search (genetic algorithm, ant colony optimization, and particle swarm optimization)29,30,31,32,33,34,35, geometric algorithms (probabilistic road map, visibility graph, rapid-exploring random tree, cell decomposition, Voronoi graph)36,37,38,39,40,41. However, selective area cleaning/spot cleaning path planning for re-configurable floor cleaning robots is a new approach. It has not been explored yet. This research describes the optimal shortest waypoint coverage path planning method using evolutionary-based optimization technique for re-configurable floor cleaning robot hTetro to execute the selective area cleaning/spot cleaning task.
Figure 1 shows the overview of our selective area cleaning/spot cleaning framework with a functional block diagram. It is developed for floor cleaning robot indoor selective area cleaning/spot cleaning task. Generally, high human traffic areas are more likely to accumulate more trash and foot stains. Hence, two strategies are combined to efficiently find the indoor environment’s selective area cleaning/spot cleaning regions. It includes tracing the human traffic patterns and detection of stains, foot stains, and trash on the floor.
The proposed system functional block diagram is composed of an image data collection module from the RGB-D vision sensor network, deep learning-based computer vision algorithms, and an optimal path planning module for the hTetro robot. WiFi-enabled surveillance camera setup (Realsense D345 fitted in roof corners) captures human movements indoors and continuously captures the floor images. The camera system is fixed in the corners of the roof, and the Field of View (FoV) is calibrated with the 3D map built by the hTetro robot base frame. The cameras are connected to an external server through a WiFi router to transfer the captured video stream’s. A deepSORT human tracking algorithm was used to trace the human traffic pattern. SSD Mobilenet object detection algorithm was trained to detect foot stains on the high human traffic region, common stains, and trash on the floor. The selective area cleaning/spot cleaning algorithm is run on an external server built with a high-performance GPU. The server collects the indoor images continuously through a WiFi router and performs the human traffic pattern detection analysis, stains, and trash detection task. According to the human traffic pattern tracing, stains, and trash detected on the floor, the framework generates the navigation strategy and send the way point coordinates to hTetro robot to perform the selective area cleaning/spot cleaning task. The robot uses waypoint information (computed through image coordinate to robot footprint coordinate) and odometry data to reach the selective area cleaning/spot cleaning location and perform the cleaning task. The details of each component and integration methods are described as follows.
Overview of the proposed system.
The Deep SORT is an improved version of the Simple Online and Real-time Tracking (SORT) algorithm. Generally, deep sort is used for human tracking applications. In our work, Deep SORT is used to monitor the human traffic pattern tracing in indoors through RGB-D vision sensor collected image and map the high traffic area to identify the selective area cleaning/spot cleaning region. The Deep SORT tracking framework was built using the hypothesis tracking technique with the Kalman filtering algorithm and DL-based association metric approach. Further, the Hungarian algorithm is utilized to resolve the uncertainty between the estimated Kalman state and the newly received measuring value. The tracking algorithm uses the appearance data to improve the performance of sort42. In this work, the bounding box coordinates in image plane (X(_{min}),Y(_{min}) and X(_{max}) and Y(_{max})) of human detection are passed to Deep SORT with image frame for tracking the human movements. According to bounding box coordinates and object appearance, deep SORT assigns each human detection ID and performs the tracking. A selective area cleaning/spot cleaning algorithm used the tracked information to compute the cleaning spot.
A Single Shot Detector (SSD) MobileNet V2 end-to-end object detection framework is adopted to detect the stains, foot stains, and trash on the floor. The functional block diagram of SSD MobileNet is shown in Fig. 2. Here, feature extraction is done using the MobileNet model. Bounding box prediction is made using a meta-architecture SSD. While training, RMSprop algorithm43 is used for loss optimization. The framework was trained using a transfer learning scheme; the last layer of the detector head is modified to detect the stains, foot stains, and trash.
SSD MobileNet V2.
MobileNet v2 is used as the base network that performs feature extraction tasks suitable for mobile and edge devices. Figure 2 shows the MobileNet v2 architecture. It comprises multiple residual bottleneck layers. These layers use (3times 3) in depthwise convolution layers, which are more efficient than standard convolution layers. They also employ (1times 1) convolutions instead. The architecture also uses ReLU6 layers, which are ReLU layers with an upper limit of 6. The upper limit prevents the outputs from scaling to a huge value, reducing the computation cost.
Single-Shot Detector (SSD) is an object detection algorithm. It outputs the object’s location through a bounding box in an image with the confidence of the prediction. The bounding box coordinates have been translated into real-world coordinates and fed to a trajectory planning algorithm to generate the selective area cleaning/spot cleaning path. The SSD algorithm uses the MobileNet feature map to generate the anchor boxes. The final few layers (feature map) from the feature extractor are used to predict trained objects’ locations. In the SSD architecture, the input is passed through different convolution layers of different sizes. These layers decrease in size progressively through the network. The purpose of these layers is to enable the network to detect objects of different shapes and sizes.
While training, confidence loss (L_{confidence}) and localization loss (L_{location}) Eq. (1) was calculated for each prediction. The prediction error of class and confidence is given by confidence loss. The localization loss is a smooth loss between the predicted bounding box value and the ground truth value. The value (alpha ) is used to reduce the overall loss’s influence and balance confidence and location loss.
Root Mean Squared gradient descent algorithm (RMS) is used for optimizing both the losses. The weights (w_t) are computed at any time t using the gradient of loss L, (g_t) and gradient of the gradient (v_t) Eqs. (2)–(4). The hyperparameter (beta = 0.9, eta = 0.02) are utilized for estimating the momentum and gradient values, while a small value near to zero (epsilon ) is used for avoiding divide-by-zero errors.
After finding the 2D pixel coordinates in the RGB image frame of stain, trash, and human traffic pattern, the system will derive the world frame’s corresponding 3D position. Since, the perception system is a stereo camera with both an RGB image sensor and a depth image sensor, they are of the same resolution, and each pixel in the RGB image corresponds to the pixel in the depth image. This enables us to map the dirt object’s localized pixel coordinates to their depth value. When the depth images are subscribed, the image will be filtered via an adaptive directional filter44 to remove noise. After filtering the depth image, the depth value of the dirt objects, (d_o), has been derived by Eq. (5) as follows:
where, (Omega ) represents the sum of pixels within a (10 times 10) window, and (d_i) represents the value of the depth in the pixel. The dirt objects in real-world coordinates will be estimated by converting the X and Y pixel coordinate system to x and y world coordinate system built by the Robot Lidar sensor.
RGB-D Sensor parameters are derived through camera calibration techniques to give the camera an intrinsic and extrinsic matrix45. Camera calibration methods give us the focal length, (f_x), and (f_y), of the RealSense RGB-D sensor and the optical center, (c_x), and (c_y), of the RealSense RGB-D sensor in its respective x and y coordinate to get the intrinsic camera matrix, I. The RealSense RGB-D sensor in-built functions also provide us with the translation vector, K, and rotation vector, Q. The extrinsic matrix, S, can be derived where (S = [Q K]). Further, it will convert from the pixel coordinate system to the world coordinate system through Equation (6).
where P is the world coordinates, p is the pixel coordinates.
hTetro is an in-house developed re-configurable floor cleaning robot (Fig. 4). This robot consists of a chain-type with four cube shape modules connected by three hinges at the corner of blocks, as shown in Fig. 3. In hTetro, blocks 1, 3, and 4 each hold a cleaning module, while block 2 contains the primary electronic circuits. The robot blocks could be easy to remove and add a degree of freedom. The hinges permit the hTetro robot in shape-transferring and a couple of configurations. The shape-shifting allows the robot to cover the constrained space more efficiently than fixed-form robots. For instance, the robot can cover the bigger space with the default O shape and change shapes flexibly from the O to I shape to cover the narrow spaces. The three hinges connected between robot blocks are used for shape-shifting during the reconfiguration.
The locomotion modules at each block are programmed to operate synchronously with others according to the robot’s transformation, and each of the modules is amenable for steering and moving the hTetro. Each block contains two geared DC motors with encoders for balance locomotion. The operation voltage rating of the DC motors is 7.4V. In hTetro, a potentiometer type absolute encoder was used in each block to sense the shape change action and block’s heading angle. When the hTetro is involved in the transformation actions, the servos located at the hinges drive the modules by rotation clockwise and anticlockwise to do the reconfiguration into seven distinct shapes –‘I, L, S, T, Z, J, O’, as shown in Fig. 3. The low-level controller (Arduino mega) was used to measure the output current from the absolute encoder through the A/D Input pin and fine-tunes the shape change action. This sensing data helps ensure that the units are aligned and moving in the same direction. The servo motors run at 14.8V, and the high torque is 77 kg. cm. The servo has enough torque to hold the hTetro block masses in morphologically changing and fastening the blocks in locomotion.
hTetro hardware architecture was developed with a single-board computer Intel compute stick and low-level micro-controller Arduino mega 2560 (Fig. 4). The Intel compute stick is powered with an Intel Core m3 processor, 4GB of memory, and 64GB of eMMC flash storage. It was used as an hTetro central processing unit, runs with an operating system of Ubuntu 20.04, and uses ROS Melodic46 as middleware. The hTetro navigation algorithms run alongside the ROS middleware in the compute stick and control all the autonomy-related operations of the hTetro. The navigation algorithm receives the waypoint information from the remote/local server through on board WiFi module and generates the control command to the low-level control unit to execute the selective area cleaning/spot cleaning task.
Arduino-Mega was used as a low-level controller. It works at 5 volt supply voltage and operates at 16 Mhz clock frequency. The controller receives the commands from the central processing unit through ROS serial bridge function and generates the appropriate actions for the motor drives and other local electronic components. The rosserial ROS package uses Arduino’s universal asynchronous receiver/transmitter (UART) communication and converts the board to a ROS node that can output ROS messages and subscribe to messages as well.
Seven configurations of hTetro.
hTetro hardware components.
The 3D points at the centers of detected objects are defined as waypoints the robot needs to visit to do the appropriate selective area cleaning/spot cleaning region. The robot will derive the optimal navigation trajectories defined as Complete Waypoint Coverage Planning (CWCP) to visit all the waypoints. With N unsequenced waypoints detected from the perception unit, the navigation trajectory (eta ) is defined as follows: (eta ~=~{(W_1^s, W_1^g), (W_2^s, W_2^g) …, (W_N^s, W_N^g)}) The trajectory (eta ) consists of pairs of waypoints with the directions which indicate the order of navigation. The initial waypoint is labeled (W_1^s), and the final waypoint is labeled (W_N^g). This trajectory representation visits each waypoint among N waypoints once. The cost to navigate from source waypoint to destination waypoint equals to summation of mass of each robot (m_i) block multiplied with 3D displacement to move blocks from source ((x^s_i,y^s_i,z^s_i)) to goal ((x^g_i,y^g_i,z^g_i)) as shown in Eq. (7). The total cost of the trajectory can be summarized by all the waypoint pair components as shown in Eq. (8). The Pareto trajectory is defined as (widehat{eta }), and the objective of the proposed CWPP problem is derived by Eq. (9).
There are many possibilities to form the trajectories and find the optimal route to minimize the fitness function; the classic Travelling Salesman Problem (TSP) model can be considered an NP-hard problem. Similarly, brute force search can be used for iterating every possible path with factorial time complexity O(n!), But performance gradually decreases with the large input data size. Dynamic programming through Held-Karp algorithm47 reduces the complexity of TSP to O((n^22^n)) time. However, the method is still inappropriate for the presented scenario where multiple detected waypoints are in the working workspace. Evolutionary approach Ant Colony Optimization (ACO)48,49 is adopted to find the optimal trajectory connecting the waypoints, (hat{eta }) in Eq. (9) within a reasonable time. Optimization based on probabilistic techniques are the main idea of ACO. By tuning the ant colony decisions at the waypoint and the constant updates on the pheromones left along the temporary trajectory, the ACO algorithm has proven to be a feasible method to derive the Pareto-optimal solution of the TSP while eliminating the calculation burden.
This section presents the experimental procedure and outcome of the selective area cleaning/spot cleaning framework tested on the hTetro robot. The algorithm has two pre-processing steps: training the trash and stain detection framework using a custom dataset and mapping the indoor environment for hTetro autonomous navigation.
The dataset was prepared in-house with different image resolutions. The dataset has four classes (Stain, Foot stain, Trash and Human); each class has 1200 images. Most of the images are created in-house with various floor backgrounds, and few images are collected online through bing image search. Here, in-house collected images are captured in an overhead perspective with a Realsense RGBD camera with a different angle and different illumination conditions. Then, the captured images are labeled for training the detection framework (SSD MobileNet). Further, to reduce the CNN learning rate and avoid over-fitting. The data augmentation (rotation, scaling, and flipping) is applied in collected images.
The standard method of the K-fold (here K = 10) cross-validation process is used to evaluate the detection model. In K-fold cross-validation, the labeled images are split into ten sets, and nine sets were used to train the detection network, and one set is utilized for evaluation. These steps run for ten rounds to ensure all the data were used for both training and testing to eliminate the biasing conditions. This process is done to remove any bias due to the particular training and testing data set to split. The experimental results are obtained from the detection model with reasonable accuracy. The deep learning models were developed in Tensor-flow 1.9 Ubuntu 18.04 version and trained using the following hardware configuration Intel core i7-8700k, 64 GB RAM, and Nvidia GeForce GTX 1080 Ti Graphics Cards. Standard statistical measures such as accuracy (10), precision (11), recall (12) and (F_{measure}) (13) were used to evaluate the proposed algorithm through confusion matrix elements.
Here, tp, fp, tn, fn represents the true positives, false positives, true negatives, and false negatives, respectively, as per the standard confusion matrix.
The selective area cleaning/spot cleaning framework was validated in three different indoor environments with the sizes of (10 times 10, 8 times 8, 15 times 8 ) feet. The overhead cameras were fixed corner of the roof in the testbed for tracing the human movements and capturing the floor images. The captured floor images are transferred to the external server, running a selective area cleaning/spot cleaning framework. The experiment was carried out with different lighting conditions, and results are shown in Figs. 5 and 6. In this trial, the overhead camera run at 30fps, and the image resolution was (640times 480) pixels. The floor image was captured at 7 feet height from the ground, and the camera head is inclined at an angle of 45 degrees downwards.
Real-time detection of stain, foot stain, trash.
Human movement tracking, (a–c) human movements and (d) tracking result.
The stain is detected by a light turquoise color bound box, foot stains were marked as white rectangle box and trash marked by the sky blue bounding box. Figures 5 and  6 shows the human traffic pattern detection results. Here, the human movements were tracked by bounding boxes, and the traffic pattern is indicated in the orange line. The experiment results indicate that the selective/ spot cleaning framework detects the human, stain, and trash with a higher confidence level of (92%), (90%), 95%, and (98%), respectively. Further, the model detection accuracy was estimated by statistical measures, which are given in Table 2.
The statistical measurement result shows that the framework has identified the stain with an average of 93.44%, foot stains with 91.69%, trash with 93.0 %, and human with 97.0%, respectively. In this experiment, we observe that stain and foot stain detection confidence level and statistical measure values are relatively low compared to other classes due to less visibility of objects and complex texture shape. Further, the computation cost of the proposed system was estimated via inference time; in this test, the framework took 0.05 seconds to process the (640 times 480) images. The statistical study ensures that the proposed system was stable for different lighting conditions and also work in realtime.
The efficiency of the SSD mobilenet stain trash detection was evaluated with other benchmark object detection algorithms include Faster RCNN ResNet50, Faster RCNN Inception51, and Yolo v2 frameworks. The same dataset was used for estimating the detection and classification accuracy of three networks for a period at the same time. Statistical measure parameter was used to compare the performance model. Table 3 shows statistical measure results of each framework. In addition to that, timing analysis was performed to calculate the detection time taken by each network. NVIDIA Quadro P4000 graphic card was used for testing the proposed algorithm, as shown in Table 4.
Best detection accuracy was obtained using Faster RCNN ResNet and Faster RCNN inception compared to other models. Yolo v2 inference time is comparatively low but showed poor detection accuracy. SSD-MobileNet has obtained a better trade-off between detection accuracy and inference time. Hence, SSD- MobileNet shows optimum results when used in realtime applications for object detection. Optimized SSD MobileNet is highly suitable for low-power resource-constrained embedded systems, making it suitable for realtime applications.
We validated the efficiency and performance of the selective area cleaning/spot cleaning method in terms of navigation efficiency by deploying the cleaning robot system to clean optimally the detected tracked human traces in the selected indoor testbed sites. Specifically, depending on the human interaction activity in the environment, a location map is constructed on the grid cells by the proposed framework. The location map distribution result is used as input for a hTetro reconfigurable cleaning robot to decide navigation strategies optimally.
The Complete Waypoints Coverage Planning (CWCP) framework derives the smallest costweight and energy consumption for real environments testbed. The trajectories denoted as red linking arrows generated by all teed methods including zigzag, spiral, greedy search with 100 iterations, and ACO algorithms for all testbed workspace-1 with 9 detected waypoints and workspace-2 with 11 detected waypoints are shown in Figs. 7 and 8, respectively. In the experiment, the best parameter values are found by the experimental approach by 200 trials. after tunning, the ACO parameters set is evaporation probability = 0.9, and ant agents = 220. The terminal’s criteria are triggered when the costweight during execution is not improved for ten iterations or the execution loop runs over the 500 iterations.
The trajectories of the tested method for workspace-1. (a) zigzag; (b) spiral; (c) greedy search; (d) ACO.
The trajectories of the tested method for workspace-2. (a) zigzag; (b) spiral; (c) greedy search; (d) ACO.
The trajectories with found costweight and the time to generate as well as the running time to complete these found trajectories of the tested method for the two testbed workspaces are represented in Table  5, respectively. Because of the straightforward approach in generating the paths, the path generation time of the zigzag and the spiral are smaller than Greedy Search and ACO, but the zigzag and spiral methods link the defined locations by position-wise orders; hence the costweight is greater than greedy search methods and significantly greater than ACO based Complete Waypoints Path Planning (CWPP). Noted that a greedy algorithm is an algorithm that follows the problem-solving heuristic of making the locally optimal choice at each stage. Since it may be stuck at local minimal so we run the greedy algorithm for 100 iteration with random beginning waypoint and select the best solution. That makes that greed search requires more path generation time.
In terms of running time when deploying the found trajectory to the real robot, the Zigzag and Spiral need more time to cover the defined waypoint than the greedy search. Together with the least costweight, the running time of ACO-based CWPP is the shortest one.
After the perception system has defined the waypoint locations in real testbed sites, the energy efficiency during the selective area cleaning/spot cleaning process is validated by recording in realtime the energy consumption. The current sensors connected to the main lithium-ion battery with the specs of 14.4 V, 3000 mAh are logged during robot operation. These current values are integrated with the running time to find power consumption. In the selective area cleaning/spot cleaning mode, the hTetro is set at O shape and to clear these waypoints in order. The robot can change to I shape to access the narrow space area by the environment evaluation from the Lidar sensor. The action of shape-changing is adapted to a dynamic environment. The robot modules such as perception, locomotion, and shape-shifting communicate under the controls by the ROS system52. To archive the smooth locomotion with the different kinematic design of hTetro, the moving command generated by the PID controller is sent to motor drivers to issue appropriate velocity for the DC motors. Specifically, the robot should keep the planned trajectories and overcome the disturbances while colliding with static and dynamic obstacles. These tasks are significant challenges since the synchronization between hTetro blocks locomotion must be accomplished. To fulfill the kinematic constraints, the control method is based on the Instantaneous Center of Rotation (ICR). The placement of ICR makes use of the desired positions and velocities information provided by the central computing unit based stains pattern. After that, the corresponding steering angle and required individual velocity are calculated based on the ICR location. A velocity regulator is then used to limit the velocity of each motor. Due to the placement of LIDAR on the second module’s chassis, the orientation of the whole robot is defined according to it. The instantaneous center of rotation’s desired location is defined based on the desired position, current position, and the desired velocity. It is represented in the polar coordinate using to avoid numerical singularity. The hTetro locates itself within the pre-built map using enhanced extended Kalman filter(EKF)53 for multi-sensor fusion makes sure hTetro can navigate from source and destination goad correctly even if sensors noise is received, or any perception sensors struggle the malfunction issues.
From Table  5, if the robot clears the found path of the method with a smaller costweight value, the less energy spent is observed. Specifically, since producing the longest costweight, the zigzag draws the highest energy, then the spiral follows the following position with a slight amount lower. The best CWCP method in both energy and time saving is ACO. It yields about 10% and 15% lower than the greedy search in energy and time spent, respectively. The results ensure that the proposed CWCP with ACO is possible for selective area cleaning/spot cleaning by hTetro.
This work proposed a selective area cleaning/spot cleaning technique for reconfigurable floor cleaning robot hTetro using deep learning-based computer vision algorithms and an optimal complete waypoints path planning method. SSD MobileNet object detection algorithm and Deep SORT real-time human tracking algorithm were used to find the human traffic patterns and detect the stains, foot stains, and trash on the floor. The proposed scheme was tested with two different testbeds with hTetro reconfigurable floor cleaning robot and evolutionary optimization based optimal trajectory generation algorithm. The experimental results showed that the trained SSD model detects the stains, foot stains, and trash on the floor with a high confidence score, and its result is used to find the selective area cleaning/spot cleaning region. Further, the SSD MobileNet framework’s stain and trash detection efficiency was evaluated with other CNN-based object detection algorithms and observed that SSD MobileNet obtained a better trade-off between detection accuracy and low inference time for detection of stains and trash. The overall experimental results show that the proposed framework is more suitable for enabling the autonomous selective area cleaning/spot cleaning function for the reconfigurable floor cleaning robot hTetro.
‘The datasets used and/or analysed during the current study available from the corresponding author on reasonable request.’ Correspondence and requests for materials should be addressed to B.R.
The 5 most unwanted jobs in singapore, accessed 6 Jan 2020; https://sg.finance.yahoo.com/news/5-most-unwanted-jobs-singapore-160000715.html.
Yin, J. et al. Table cleaning task by human support robot using deep learning technique. Sensors 20, 5. https://doi.org/10.3390/s20061698 (2020).
Article  Google Scholar 
Murdan, A. P. & Ramkissoon, P. K. A smart autonomous floor cleaner with an android-based controller. In 2020 3rd International Conference on Emerging Trends in Electrical, Electronic and Communications Engineering (ELECOM), 235–239, https://doi.org/10.1109/ELECOM49001.2020.9297006 (2020).
Deepa, G. et al. Robovac-automatic floor cleaning robot. In 2021 Third International Conference on Inventive Research in Computing Applications (ICIRCA), 1–4, https://doi.org/10.1109/ICIRCA51532.2021.9544914 (2021).
Ramalingam, B. et al. Deep learning based pavement inspection using self-reconfigurable robot. Sensors 21, 5. https://doi.org/10.3390/s21082595 (2021).
Article  Google Scholar 
Ramalingam, B., Veerajagadheswar, P., Ilyas, M., Elara, M. R. & Manimuthu, A. Vision-Based Dirt Detection and Adaptive Tiling Scheme for Selective Area Coverage. Journal of Sensors. https://doi.org/10.1155/2018/3035128 (2018).
Ramalingam, B., Lakshmanan, A.K., Ilyas, M., Le, A.V. & Elara, M.R. Cascaded machine-learning technique for debris classification in floor-cleaning robot application (2018).
Shin, M., Paik, W., Kim, B. & Hwang, S. An iot platform with monitoring robot applying cnn-based context-aware learning. Sensors 19, 2525. https://doi.org/10.3390/s19112525 (2019).
ADS  Article  PubMed Central  Google Scholar 
Koo, J. H., Cho, S. W., Baek, N. R., Kim, M. C. & Park, K. R. Cnn-based multimodal human recognition in surveillance environments. Sensors 18, 3040. https://doi.org/10.3390/s18093040 (2018).
ADS  Article  PubMed Central  Google Scholar 
Andersen, N.A., Braithwaite, I.D., Blanke, M. & Sorensen, T. Combining a novel computer vision sensor with a cleaning robot to achieve autonomous pig house cleaning. In Decision and Control, 2005 and 2005 European Control Conference. CDC-ECC’05. 44th IEEE Conference on 8331–8336 (IEEE, 2005).
Bormann, R., Fischer, J., Arbeiter, G., Weisshardt, F. & Verl, A. A visual dirt detection system for mobile service robots. In ROBOTIK 2012; 7th German Conference on Robotics 1–6 (2012).
Bormann, R., Weisshardt, F., Arbeiter, G. & Fischer, J. Autonomous dirt detection for cleaning in office environments. In 2013 IEEE International Conference on Robotics and Automation 1260–1267 (2013).
Milinda, H.G.T. & Madhusanka, B.G.D.A. Mud and dirt separation method for floor cleaning robot. In 2017 Moratuwa Engineering Research Conference (MERCon) 316–320 (2017).
Lee, H. & Banerjee, A. Intelligent scheduling and motion control for household vacuum cleaning robot system using simulation based optimization. In 2015 Winter Simulation Conference (WSC) 1163–1171 (2015).
Grünauer, A., Halmetschlager-Funek, G., Prankl, J. & Vincze, M. The power of gmms: Unsupervised dirt spot detection for industrial floor cleaning robots. In (eds. Gao, Y. et al.) Towards Autonomous Robotic Systems 436–449 (Springer International Publishing, 2017).
Martínez, D., Alenyà, G. & Torras, C. Planning robot manipulation to clean planar surfaces. In Engineering Applications of Artificial Intelligence 39 23–32. https://doi.org/10.1016/j.engappai.2014.11.004 (2015).
Article  Google Scholar 
Park, J., Chen, J., Cho, Y. K., Kang, D. Y. & Son, B. J. Cnn-based person detection using infrared images for night-time intrusion warning systems. Sensors 20, 34. https://doi.org/10.3390/s20010034 (2020).
ADS  Article  Google Scholar 
Gomaa, A., Abdelwahab, M. M., Abo-Zahhad, M., Minematsu, T. & Taniguchi, R.-I. Robust vehicle detection and counting algorithm employing a convolution neural network and optical flow. Sensors 19, 4588. https://doi.org/10.3390/s19204588 (2019).
ADS  Article  PubMed Central  Google Scholar 
Fulton, M., Hong, J., Islam, M.J. & Sattar, J. In Robotic Detection of Marine Litter Using Deep Visual Detection Models. arXiv:1804.01079 (2018).
Rad, M.S. et al. A computer vision system to localize and classify wastes on the streets. In ICVS (2017).
Zhihong, C., Hebin, Z., Yanbo, W., Binyan, L. & Yu, L. A vision-based robotic grasping system using deep learning for garbage sorting. In Control Conference (CCC), 2017 36th Chinese 11223–11226 (IEEE, 2017).
Cauli, N. et al. Autonomous table-cleaning from kinesthetic demonstrations using deep learning. In 2018 Joint IEEE 8th International Conference on Development and Learning and Epigenetic Robotics (ICDL-EpiRob) 26–32 (2018).
Farooq, A., Farooq, F. & Le, A. V. Human action recognition via depth maps body parts of action. TIIS 12, 2327–2347 (2018).
Google Scholar 
Lakshmanan, A. K. et al. Complete coverage path planning using reinforcement learning for tetromino based cleaning and maintenance robot. Autom. Constr. 112, 103078 (2020).
Article  Google Scholar 
Dijkstra, E. W. A note on two problems in connexion with graphs. Numer. Math. 1, 269–271. https://doi.org/10.1007/BF01386390 (1959).
MathSciNet  Article  MATH  Google Scholar 
Hart, P. E., Nilsson, N. J. & Raphael, B. A formal basis for the heuristic determination of minimum cost paths. IEEE Trans. Syst. Sci. Cybern. 4, 100–107. https://doi.org/10.1109/TSSC.1968.300136 (1968).
Article  Google Scholar 
Stentz, A. Optimal and efficient path planning for partially-known environments. In Proceedings of the International Conference on Robotics and Automation 3310–3317. (1993).
Koenig, S. & Likhachev, M. Fast replanning for navigation in unknown terrain. IEEE Trans. Robot. 21, 354–363. https://doi.org/10.1109/TRO.2004.838026 (2005).
Article  Google Scholar 
Nazarahari, M., Khanmirza, E. & Doostie, S. Multi-objective multi-robot path planning in continuous environment using an enhanced genetic algorithm. Expert Syst. Appl. 115, 106–120. https://doi.org/10.1016/j.eswa.2018.08.008 (2019).
Article  Google Scholar 
Cheng, K. P., Mohan, R. E., Nhan, N. H. K. & Le, A. V. Multi-objective genetic algorithm-based autonomous path planning for hinged-tetro reconfigurable tiling robot. IEEE Access 8, 121267–121284. https://doi.org/10.1109/ACCESS.2020.3006579 (2020).
Article  Google Scholar 
Patle, B. K., Parhi, D. R. K., Jagadeesh, A. & KumarKashyap, S. Matrix-binary codes based genetic algorithm for path planning of mobile robot. Comput. Electr. Eng. 67, 708–728. https://doi.org/10.1016/j.compeleceng.2017.12.011 (2018).
Article  Google Scholar 
Thabit, S. & Mohades, A. Multi-robot path planning based on multi-objective particle swarm optimization. IEEE Access 7, 2138–2147. https://doi.org/10.1109/ACCESS.2018.2886245 (2018).
Article  Google Scholar 
Das, P. K., Behera, H. S. & Panigrahi, B. K. A hybridization of an improved particle swarm optimization and gravitational search algorithm for multi-robot path planning. Swarm Evol. Comput. 28, 14–28. https://doi.org/10.1016/j.swevo.2015.10.011 (2016).
Article  Google Scholar 
PortaGarcia, M., OscarMontiel, Castillo, O., Sepúlveda, R. & Melin, P. Path planning for autonomous mobile robot navigation with ant colony optimization and fuzzy cost function evaluation. Appl. Soft Comput. 9, 1102–1110. https://doi.org/10.1016/j.asoc.2009.02.014 (2009).
Article  Google Scholar 
Zeng, M.-R., Xi, L. & Xiao, A.-M. The free step length ant colony algorithm in mobile robot path planning. Adv. Robot. 30, 1509–1514. https://doi.org/10.1080/01691864.2016.1240627 (2016).
Article  Google Scholar 
Clark, M. C. Probabilistic road map sampling strategies for multi-robot motion planning. Robot. Auton. Syst. 53, 244–264. https://doi.org/10.1016/j.robot.2005.09.002 (2005).
Article  Google Scholar 
Niewola, A. & Podsedkowski, L. L* algorithm-a linear computational complexity graph searching algorithm for path planning. J. Intell. Robot. Syst. 91, 425–444. https://doi.org/10.1007/s10846-017-0748-6 (2018).
Article  Google Scholar 
Karaman, S. & Frazzoli, E. Sampling-based algorithms for optimal motion planning. Int. J. Robot. Res. 30, 846–894. https://doi.org/10.1177/0278364911406761 (2011).
Article  MATH  Google Scholar 
Qureshi, A. H. & Ayaz, Y. Intelligent bidirectional rapidly-exploring random trees for optimal motion planning in complex cluttered environments. Robot. Auton. Syst. 68, 1–11. https://doi.org/10.1016/j.robot.2015.02.007 (2015).
Article  Google Scholar 
Li, Y., Chen, H., Er, M. J. & Wang, X. Coverage path planning for uavs based on enhanced exact cellular decomposition method. Mechatronics 21, 876–885. https://doi.org/10.1016/j.mechatronics.2010.10.009 (2011).
Article  Google Scholar 
Candeloro, M., Lekkas, A. M. & Sørensen, J. A. A voronoi-diagram-based dynamic path-planning system for underactuated marine vessels. Control Eng. Pract. 61, 41–54. https://doi.org/10.1016/j.conengprac.2017.01.007 (2017).
Article  Google Scholar 
Bewley, A., Ge, Z., Ott, L., Ramos, F. & Upcroft, B. Simple online and realtime tracking. In 2016 IEEE International Conference on Image Processing (ICIP) 3464–3468 (2016).
Tieleman, T. & Hinton, G. Lecture 6.5-rmsprop, coursera: Neural networks for machine learning. University of Toronto, Technical Report (2012).
Le, A. V., Jung, S.-W. & Won, C. S. Directional joint bilateral filter for depth images. Sensors 14, 11362–11378 (2014).
ADS  Article  Google Scholar 
Bradski, G. The OpenCV Library. In Dr. Dobb’s Journal of Software Tools (2000).
Le, A. V. & Choi, J. Robust tracking occluded human in group by perception sensors network system. J. Intell. Robot. Syst. 90, 349–361 (2018).
Article  Google Scholar 
Malandraki, C. & Dial, R. B. A restricted dynamic programming heuristic algorithm for the time dependent traveling salesman problem. Eur. J. Oper. Res. 90, 45–55 (1996).
Article  Google Scholar 
Dorigo, M. & Di Caro, G. Ant colony optimization: A new meta-heuristic. In Proceedings of the 1999 congress on evolutionary computation-CEC99 (Cat. No. 99TH8406), vol. 2, 1470–1477 (IEEE, 1999).
Le, A. V., Nhan, N. H. K. & Mohan, R. E. Evolutionary algorithm-based complete coverage path planning for tetriamond tiling robots. Sensors 20, 445 (2020).
ADS  Article  Google Scholar 
Targ, S., Almeida, D. & Lyman, K. Resnet in resnet: Generalizing residual architectures. arXiv:1603.08029 (2016).
Szegedy, C., Ioffe, S., Vanhoucke, V. & Alemi, A. Inception-v4, inception-resnet and the impact of residual connections on learning. arXiv:1602.07261 (2016).
Quigley, M. et al. Ros: An open-source robot operating system. In ICRA workshop on open source software, vol. 3, 5 (Kobe, Japan, 2009).
González, J. et al. Mobile robot localization based on ultra-wide-band ranging: A particle filter approach. Robot. Auton. Syst. 57, 496–507 (2009).
Article  Google Scholar 
Download references
This research is supported by the National Robotics Programme under its Robotics Enabling Capabilities and Technologies (Funding Agency Project No. 192 25 00051), National Robotics Programme under its Robot Domain Specific (Funding Agency Project No. 192 22 00058), National Robotics Programme under its Robotics Domain Specific (Funding Agency Project No. 192 22 00108), and administered by the Agency for Science, Technology and Research.
ROAR Lab, Engineering Product Development, Singapore University of Technology and Design, Singapore, 487372, Singapore
Balakrishnan Ramalingam, Anh Vu Le, Rajesh Elara Mohan & Sathian Pookkuttath
School of Electrical and Electronics Engineering, Nanyang Technological University, Singapore, 639798, Singapore
Zhiping Lin & Zhenyu Weng
You can also search for this author in PubMed Google Scholar
You can also search for this author in PubMed Google Scholar
You can also search for this author in PubMed Google Scholar
You can also search for this author in PubMed Google Scholar
You can also search for this author in PubMed Google Scholar
You can also search for this author in PubMed Google Scholar
Conceptualization, R.E.M.; Methodology, R.E.M, Z.L.; Software, B.R, A.L., S.P., Z.W; Validation, R.E.M, Z.L.; Formal analysis, R.E.M, Z.L, B.R., S.P.; Investigation, B.R., A.L.; Resources, M.R.E.; Data curation, S.P., Z.W., Z.L; Writing—original draft, B.R., A.L , Z.W; Writing—review & editing, B.R., Z.W, Z.L; Visualization, S.P.; Supervision, M.R.E, Z.L; Project administration, M.R.E.; Funding acquisition, M.R.E. All authors have read and agreed to the published version of the manuscript.
Correspondence to Balakrishnan Ramalingam.
The authors declare no competing interests
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.
Reprints and Permissions
Ramalingam, B., Le, A.V., Lin, Z. et al. Optimal selective floor cleaning using deep learning algorithms and reconfigurable robot hTetro. Sci Rep 12, 15938 (2022). https://doi.org/10.1038/s41598-022-19249-7
Download citation
Received:
Accepted:
Published:
DOI: https://doi.org/10.1038/s41598-022-19249-7
Anyone you share the following link with will be able to read this content:
Sorry, a shareable link is not currently available for this article.

Provided by the Springer Nature SharedIt content-sharing initiative
By submitting a comment you agree to abide by our Terms and Community Guidelines. If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.
Advertisement
Advanced search
Scientific Reports (Sci Rep) ISSN 2045-2322 (online)
© 2022 Springer Nature Limited
Sign up for the Nature Briefing newsletter — what matters in science, free to your inbox daily.

source

Death of a Master Forger – History Today

October 2022
Subscription
Offers

Give a Gift
Subscribe
John Payne Collier died on 17 September 1883, after a lifetime of creating deliberate fictions, falsehoods and forgeries.

N.E.S.A. Hamilton. "An inquiry into the genuineness of the manuscript corrections in Mr. J. Payne Collier’s annotated Shakspere, folio, 1632: and of certain Shaksperian documents likewise published by Mr. Collier." Special Collections, University of Delaware Library.
‘An inquiry into the genuineness of the manuscript corrections in Mr. J. Payne Collier’s annotated Shakspere, folio, 1632: and of certain Shaksperian documents likewise published by Mr. Collier’, by N.E.S.A. Hamilton, 1860. Special Collections, University of Delaware Library. 

John Payne Collier: three words sure to chill the heart of any early modern English literary scholar. Why? Because Collier was that most interesting of phenomena: a fine scholar who was also a first-class fraud. 
Aside from including deliberate fictions and falsehoods in printed records of archival material, he also introduced forgeries into the archives themselves, faking official documents, adding information to letters and diaries, falsifying registers and inventories and more.
While connoisseurs of Collier’s forgeries can honestly disagree about which has been the most damaging, no one can argue about the extent of these impostures. The so-called Perkins Folio, in which he passed off his own emendations to Shakespeare as the work of a near contemporary, contains over 20,000 corrections. Collier lived a long life and was immensely prolific; his bibliography runs to nearly 300 pages.
To paraphrase Dryden on Ben Jonson, Collier’s footprints are everywhere in the snow of the English Renaissance. (A contemporary, less kindly, called him ‘the great literary slug … What wonder if we shall still be able to trace his slime.’)
With hindsight, it is easy to see how brazen he was. ‘This part of the work will at least have the merit of novelty and originality’, he writes of the 34 ballads of his own invention he presented as Elizabethan survivals in his 1848-49 editions of Extracts from the Registers of the Stationers Company. He could not help editorialising about the quality of his work: ‘The allegory is extremely well sustained, and the ballad must have been written by no inferior hand. It would be vain now to attempt to ascertain the authorship.’
But who is to say Collier hasn’t had the last laugh? His ballads made their way into anthologies of folk song and out into the world. His legacy, because it is so contested, is the subject of much more ongoing scholarship than that of his rivals. The authoritative account of his work as both a critic and forger, published in 2004, runs to 1,483 pages. So many footprints. So much snow.
© Copyright 2022 History Today Ltd. Company no. 1556332.

source

Final plans approved for apartment complex at Fox Valley Mall in Aurora – Chicago Tribune

This rendering shows the four-story apartment building to be built at Fox Valley Mall in Aurora along the relocated Ring Road. (Centennial / HANDOUT)
Owners of Fox Valley Mall in Aurora are set to break ground on both a luxury apartment complex of more than 300 units and a senior living center of more than 200 units.
Centennial, the real estate investment firm that owns most of the mall property, this week received approval from the city of Aurora of final plans for the 323-unit, four-story apartment building on the north end of the mall.
They received approval of the final plans for the senior living center several months ago. Building is already underway – and almost finished – on the first apartments on the mall property on the eastern end of the mall where the Sears store used to be.
According to a news release from Centennial, officials plan to break ground on both the luxury apartments and the senior living center in early 2023, with each expected to open in late 2024.
Expected to open yet this month is a new green space – known as Fox Valley Yard – that would be woven between all the new residential development and the retail and entertainment uses in the mall building itself.
The one-acre community green space will host events and entertainment for both residents of the new apartments and shoppers at the mall. It will include a landscaped grassy area, a half-mile walking trail, a dog park, casual picnic-style seating and a stage for events.
Events contemplated for the space include food truck festivals, farmers markets, an outdoor beer garden, outdoor movie nights or comedy shows, school dance recitals or concerts, live art events, fitness instruction, children’s music and theater classes and pop-up boutiques.
“Centennial is well known for its creativity and its vision for the future of retail real estate,” said Whitney Livingston, president and chief operating office for Centennial. “The changes we are implementing at Fox Valley will create tremendous value for the center, the residents in the surrounding Aurora and Naperville communities, and for those people who choose to call Fox Valley their home.”
Livingston said Fox Valley is turning into “a live-work-play destination that will be unrivaled in its market.”
Scott Samson, senior general manager for the shopping center, said the redevelopment at the center is “changing how people interact with the shopping center in dramatic and imaginative ways.”
Samson added that there is “so much more to come.”
The new apartments will be the second apartment community addition to Fox Valley. The four-story, 323-unit building will include a mix of studio, one- and two-bedroom units and will offer multiple amenity spaces, outdoor gathering areas and access to neighborhood amenities.
The senior living, in a new five-story building, will offer independent living, assisted living and memory care services across 211 units.
[email protected]
Copyright © 2022, Chicago Tribune
Copyright © 2022, Chicago Tribune

source

Live coverage: Delta 4-Heavy rocket launches from California – Spaceflight Now – Spaceflight Now


Live coverage of the countdown and launch of a Delta 4-Heavy rocket with the classified NROL-91 payload for the National Reconnaissance Office. Text updates will appear automatically below. Follow us on Twitter.

United Launch Alliance’s live launch broadcast begins at 2:33 p.m. PDT (5:33 p.m. EDT; 2133 GMT) on Saturday, Sept. 24.

SpaceX launches another successful Starlink mission from Florida
NASA waves off Tuesday launch for Artemis moon rocket
Live coverage: Delta 4-Heavy rocket launches from California
Photos: Delta 4-Heavy rocket on the launch pad in California
Historic launch pad faces uncertain future after final West Coast Delta 4 mission
© 1999-2021 Spaceflight Now / Pole Star Publications Ltd

source

Spatial and temporal parasite dynamics: microhabitat preferences and infection progression of two co-infecting gyrodactylids – Parasites & Vectors – Parasites & Vectors

Advertisement
Parasites & Vectors volume 15, Article number: 336 (2022)
1 Altmetric
Metrics details
Mathematical modelling of host-parasite systems has seen tremendous developments and broad applications in theoretical and applied ecology. The current study focuses on the infection dynamics of a gyrodactylid-fish system. Previous experimental studies have explored the infrapopulation dynamics of co-infecting ectoparasites, Gyrodactylus turnbulli and G. bullatarudis, on their fish host, Poecilia reticulata, but questions remain about parasite microhabitat preferences, host survival and parasite virulence over time. Here, we use more advanced statistics and a sophisticated mathematical model to investigate these questions based on empirical data to add to our understanding of this gyrodactylid-fish system.
A rank-based multivariate Kruskal-Wallis test coupled with its post-hoc tests and graphical summaries were used to investigate the spatial and temporal parasite distribution of different gyrodactylid strains across different host populations. By adapting a multi-state Markov model that extends the standard survival models, we improved previous estimates of survival probabilities. Finally, we quantified parasite virulence of three different strains as a function of host mortality and recovery across different fish stocks and sexes.
We confirmed that the captive-bred G. turnbulli and wild G. bullatarudis strains preferred the caudal and rostral regions respectively across different fish stocks; however, the wild G. turnbulli strain changed microhabitat preference over time, indicating microhabitat preference of gyrodactylids is host and time dependent. The average time of host infection before recovery or death was between 6 and 14 days. For this gyrodactylid-fish system, a longer period of host infection led to a higher chance of host recovery. Parasite-related mortalities are host, sex and time dependent, whereas fish size is confirmed to be the key determinant of host recovery.
From existing empirical data, we provided new insights into the gyrodactylid-fish system. This study could inform the modelling of other host-parasite interactions where the entire infection history of the host is of interest by adapting multi-state Markov models. Such models are under-utilised in parasitological studies and could be expanded to estimate relevant epidemiological traits concerning parasite virulence and host survival.
In the field of theoretical and applied ecology, mathematical modelling of host-parasite systems has seen tremendous advancement and applicability [1,2,3,4]. Mathematical models provide a logical framework for developing, testing and evaluating ecological hypotheses and biological systems. These models can be categorised as individual-based models (IBMs), population-based models (PBMs) or hybrid models (i.e., integration of IBMs and PBMs) [5]. IBMs strictly model each individual by tracking the state of each member of the population, whereas PBMs track the total number of individuals in each state. Common modelling setbacks often lie with the models’ underlying assumptions being too simple or too complex [6]. For instance, many traditional ecological or epidemiological models (e.g., logistic, Lotka-Volterra predator-prey and compartmental models) assume that all individuals (within a subgroup) are identical and can be lumped together to represent the population size (as a single state variable) [7, 8]. This population lumpability means that all details regarding individual physiological and behavioural traits (determined by their specific genetic, age structure and other environmental factors) are lost. For spatially dependent systems, individuals typically affect other organisms within their spatial-temporal neighbourhood [9, 10]. Nevertheless, the data-generating processes that characterise parasite aggregation in the standard statistical frameworks are mostly not explicitly described or further explored [7]. Consequently, many studies have been conducted to bridge these modelling gaps associated with PBMs by adopting IBMs through computer simulations [11]. Incorporating evidence from the individual level to investigate processes at the population (e.g., during survival analysis), community and ecosystem levels can also improve PBMs (through hybrid models by leveraging the respective advantages of IBMs and PBMs) [12].
Through survival analysis, an individual’s infection history can be modelled as a two-stage process, with the simplest transition being from “alive” to “dead” state [13]. In such instances, we often adopt the standard logistic regression or Cox proportional-hazards regression to investigate risk factors of host mortality and to estimate hazard rates, while the non-parametric Kaplan-Meier method is used to estimate survival curves from censored data [14]. However, in most longitudinal studies, the “alive” state could further be divided into two or more intermediate (transient) phases, each corresponding to a different stage of an infection [15]. Multi-state models (MSMs) allow for time-to-event longitudinal data analysis in which surviving individuals may have different health outcomes over time. MSMs can be considered a type of hybrid model since they model several events in a given population while capturing population heterogeneity based on each individual’s infection history. A change of (infection or disease) state is considered a transition or an event. Estimating progression rates, transition probabilities, the mean sojourn time in a given state and analysing the effects of individual risk factors, survival rates and prognostic forecasting are all areas of interest under multi-state modelling [16]. For biomedical applications, clinical symptoms (such as bleeding episodes), biological markers (e.g. CD4 cell counts or serum immunoglobulin levels), scale of the disease (e.g. stages of cancer or HIV infection), or a non-fatal complication during infection progression (e.g. cancer recurrence) could all be used to classify states [16]. The states could be transient or absorbing if no other transitions could emerge from the state (e.g. after a death event). Particularly in medical settings, it is common that the exact time for some individuals to transition into an absorbing state (e.g. death) after the beginning of the study may be unknown or never quantified if occurring after the observation period. This leads to the issue of censoring in survival analysis (namely, right-censored, left-censored and interval-censored observational times), and the censoring effect must be included in multi-state modelling, especially when constructing the model’s likelihood function [16]. The complexity of an MSM in any time-to-event modelling problem depends on the number of states and all possible transitions. Hence, a more robust model (such as MSM) is useful to employ when studying host-parasite infection dynamics (including host survival and parasite virulence) to reflect the biological realism of the system under study and improve upon the estimation of risk parameters such as survival probabilities and hazard ratios [17].
Hence, multi-state models are considered a natural extension of the standard survival models [18, 19]. Andersen and Borgan [20] and Hoem et al. [21] previously reviewed Markov models, and Cox and Miller [22] broadly discussed multi-state models. In epidemiology, the states of the process could also be defined as disease outcomes such as healthy, exposed, infected, diseased with complications, or dead. The state structure (which is not unique) describes the states and the various transitions between them. For each possible transition, an MSM is specified entirely by its state structure (defined by the transition rate at which an event occurs over time) and the form of the hazard function for the respective transitions (given individuals’ characteristics or covariates). Moreover, an MSM can either be time homogeneous (if the transition intensities or rates of the Markov chain are invariant over time) or time inhomogeneous (in the case of time-varying transition rates). Also, the time-space can be either discrete or continuous. Other studies of infection diseases have employed discrete-time multi-state Markov models (e.g.,[23, 24]). Comprehensive and flexible software packages (e.g. msm and msSurv R packages) help modellers fit any proposed continuous-time multi-state Markov model (for a given biological system) based on panel or longitudinal data and well-defined transition intensities [25, 26].
In population demography, there are three broad areas for adopting MSMs [27]. The first approach is that this class of models can be used to generalise the basic life table to a more general nonhierarchical increment-decrement life table [28]. These linear models for homogeneous groups are developed using aggregate level data and are based on Markov chains. They are related to multi-state projection models, which result in asymptotic populations that are stable or temporarily stable. The second approach of MSMs (dubbed multi-state event-history models) employs event history analysis based on individual data to allow interactions between different processes and incorporate population heterogeneity [29]. This heterogeneity may be time dependent and vary according to origin or destination state. In the third approach (used for contextual and multilevel multiprocess models), both individual and aggregate measures are analysed simultaneously (in a hybrid manner). Contextual models are a straightforward extension of traditional modelling techniques, including aggregation, where the behaviours of individuals in the same context are considered independent [30]. Software like the aML package can handle fitting such models [31]. Multilevel multiprocess models are simultaneous equation systems that include multilevel hazard equations with correlated random effects (due to within-group dependence being introduced). One of the modern statistical softwares used for multilevel multiprocess modelling is gsem [32]. However, MSMs are rarely used for studying veterinary or wildlife host-parasite interactions. In the current study, for the first time, MSM is used to investigate the infection progression of two co-infecting gyrodactylids across different fish hosts.
Gyrodactylids are common fish parasites [33]. Gyrodactylus salaris alone caused epidemics among farmed salmonids, which resulted in death of up to 86% of salmon in infected rivers [34, 35]. They are monogenean ectoparasites that are ubiquitous on teleosts [36]. Amongst the well-studied Trinidadian guppy populations, gyrodactylids are the dominant parasites ((ge 42%) prevalence, 3.3 mean intensity; [37]). The prevalence of Gyrodactylus species varies spatially across watercourses (lower, mid and upper courses of the rivers or lakes) and temporally among Trinidadian populations and between host sexes [38]. Gyrodactylus prevalence is higher in female guppies, but only in lower courses [38], predominately because of fish shoaling behaviour [39, 40]. The parasites have no specific transmission stage and transfer from fish to fish occurs during host contact. Their reproductive mode is similar to that of microparasites with replication occurring directly on the host (reviewed by [34]). Their hyperviviparous nature and short generation times ((<24) h at (25^circ hbox {C}); [41]) can cause population explosions with substantial spatial and temporal variation amongst different species (e.g. [42,43,44,45,46]). Many infect the skin and fins; others occur predominantly on the gills [44, 47]. Gyrodactylus turnbulli and G. bullatarudis, which both infect the guppy (Poecilia reticulata), niche partition with G. turnbulli occurring caudally [48] and G. bullatarudis rostrally [49]. According to Harris [48], as individual host infections with G. turnbulli progress, parasites migrate from the caudal fin and body to the pectoral, pelvic, dorsal and anal fins, a migration to potentially facilitate transmission. Gyrodactylids may also move to optimise feeding, reduce competition and avoid localised immune reactions [34, 50,51,52,53,54], the scorched earth hypothesis [46]. Although the respective caudal and rostral preferences of G. turnbulli and G. bullatarudis on the host are well reported (e.g. [48, 49]), consistency over time and across different fish stocks is not.
Host survival following gyrodactylid infection was previously explored by Cable and van Oosterhout [45]. They showed that mortality of guppies differed significantly between fish stocks for each parasite strain. From their experimental study, guppies were categorised according to whether they: (i) fought off the infection, (ii) remained infected, or (iii) died while infected. The fate of these guppies was predominantly affected by fish size, such that smaller guppies were more likely to clear the infection, while larger fish either died or remained infected beyond the end of the study period [45]. Parasites and their hosts compete for survival. Such co-evolutionary interactions drive virulence originating from parasite pathogenesis and host defence [55]. Together, measures of host mortality, host resistance, host recovery, mutation, superinfection, host heterogeneity and mode of transmission all contribute to explain parasite virulence [55]. Significant heterogeneity in virulence exists between Gyrodactylus turnbulli and Gyrodactylus bullatarudis strains [45] based on host mortality, host resistance and host heterogeneity, with less emphasis on host recovery as a measure of virulence [45, 56]. Although the proportion of gyrodactylid parasite-induced causalities on different fish host has been reported (e.g. [53]), the virulence of the three gyrodactylid strains on different fish stocks has not been quantified over time while accounting for possible changes in host infection status before host mortality may occur.
The current study focuses on the spatial and temporal infection dynamics of the gyrodactylid-fish system by providing new epidemiological insights with the help of a robust time-inhomogeneous MSM, and a rank-based multivariate Kruskal-Wallis test coupled with its post-hoc tests. The time-inhomogeneous MSM is considered to analyse longitudinal survival data (instead of its time-homogeneous version) since transition intensities may naturally differ across individuals or time-varying covariates. We examine gyrodactylid microhabitat preference of three parasite strains (two strains of Gyrodactylus turnbulli and one strain of G. bullatarudis) and how these preferences vary across three different fish stocks over time based on existing experimental data. The proposed MSM is developed to improve on previous estimates of survival probabilities given fish sex, fish size, fish stock and parasite strain. We further quantify and compare the virulence (measured by rate of host mortality and recovery) over time and estimate other relevant epidemiological parameters (mean time of host to remain infected and probability of infected host to either recover or die across the covariates).
The data used here are from the experimental study of Cable and van Oosterhout [45], subsequently used as the basis of an agent-based model in van Oosterhout [53]. Briefly, cultures of three different Gyrodactylus strains were used to infect three different fish stocks: Ornamental Stock (OS), Lower Aripo River fish (LA) and Upper Aripo River fish (UA); 157 guppies in total, in a full factorial design to give nine different host-parasite combinations, with 13–22 replicates per combination. Two out of the three parasite strains were Gyrodactylus turnbulli, a laboratory-bred strain (Gt3) and a wild turnbulli strain obtained from guppies caught in the Lower Aripo River, Trinidad (Gt), whereas the third strain was G. bullatarudis, also a wild type obtained from hosts in the Lower Aripo River. Both male (68) and female (89) individually isolated guppies were used for the experiment and maintained under constant environmental conditions ((25pm 0.5 ,^circ)C; 12h light/12h dark regime). All tanks and containers were kept in a randomised block design to reduce common environmental effects.
The fish considered in the experiment were naive and thus bred under parasite-free conditions. Each fish was then infected with two parasites at time 0, and parasites were counted every 48 h over a 17-day infection period. For each fish, the number of parasites was recorded across eight different body regions (tail fin, lower body, upper body, anal fin, dorsal fin, pelvic fins, pectoral fins and head). Survival data describing the various host infection statuses (remain infected, recovered from the infection or died) over time were extracted from the empirical data for the multi-state modelling. The number of surviving fish (with or without host infection loss) and dead fish across the nine different host-parasite groups over time from days 1 to 17 is tabulated in Appendix 1.
All analyses were carried out using R version 3.6.3 [57]. Images of fish were produced in Gimp software version 2.10.12 [58] and outlined in R. Two graphical summaries of the data were produced. These are available in full in the Additional file 1: Fig. S1 and Additional file 2: Fig. S2, with examples given in Figs. 1 and 2. For the first summary (Fig. 1 and Additional file 1: Fig. S1), the shading represents the log mean intensity of parasites over surviving fish. The number of surviving fish for the nine different host-parasite groups (obtained from the fully crossed design of the three parasite strains and three different host populations) generally decreased (slowly) over time from days 1 to 17 (refer to Appendix 1). For the second graphical summary (Fig. 2 and Additional file 2: Fig. S2), the eight body regions of the fish were recategorized into four: tail, lower region (comprising of the lower body, anal fin, pelvic fins and dorsal fin), upper region (made up of the upper body and pectoral fins) and the head. This re-categorisation allowed us to visually and statistically assess any caudal-rostral preference of the three parasite strains on the three fish stocks more effectively over the study period because of low parasite numbers observed on the fish fins (anal fin, pelvic fins, dorsal fin and pectoral fin).
The movement of three different gyrodactylid parasites species/strains (Gt3, Gt and Gb) across eight host body parts (tail, lower body, anal fin, pelvic fins, dorsal fin, upper body, pectoral fins, head) of different fish stocks (OS, LA and UA stocks) at four time points. The degree of blackness indicates higher mean intensity (on log scale) over surviving fish
Mean intensities (with corresponding 95% confidence intervals) of three gyrodactylid strains (Gt3, Gt and Gb) at four main body regions (tail, lower region, upper region and head) across three fish stocks (OS, LA and UA stocks) over surviving fish and across time
The multivariate Kruskal–Wallis test (MKW) is a multivariate extension of the distribution-free univariate Kruskal-Wallis test [59]. We used it to test the null hypothesis that distribution of parasite number at the four body regions (tail, lower region, upper region and head) is equal for the different host-parasite combinations at each observed time point.
Let (Y_{ij}) be a vector of the number of parasites at the four body regions for the jth fish from the ith group (host-parasite combination), where (i=1,2,3,cdots ,9) and (j=1,2,3,cdots ,n_{i}). Let (R_{ij}) be the rank corresponding to (Y_{ij}) calculated element-wise (ties are assigned a mean rank) and ({bar{R}}_{i}=sum nolimits _{j=1}^{n_i}{frac{R_{ij}}{n_i}}) then (E({bar{R}}_{i})=m=frac{n+1}{2}) under (H_0); where (n=sum nolimits _{i=1}^{9}{n_i}) is the total number of fish ((n=157)), ({bar{R}}_{i}) is the mean rank for each ith group and ({n_i}) is the number of fish in group i. The vector (U_{i}=(bar{R}_{i1}-m,bar{R}_{i2}-m,bar{R}_{i3}-m,bar{R}_{i4}-m)^prime) denotes the average ranks for the ith group corrected for m for each variate (body regions). The pooled within-group covariance matrix is estimated as
where (R_{ij}=(R_{ij1},R_{ij2},R_{ij3},R_{ij4})^{prime }) and (1=(1,1,1,1)^{prime }). The MKW test statistic (({mathcal {W}})), given as
is approximately (asymptotically) chi-squared with (k(g-1)) degrees of freedom, where (k=4) and (g=9) [59]. After performing the MKW, the univariate Kruskal-Wallis test (UKW) was used to further compare the distribution of parasites at each of the four body regions for each parasite strain (Gt3, Gt and Gb) across the fish stocks (OS, LA and UA) at each time point (days 1 to 17). A Bonferroni-Dunn’s post-hoc test was finally applied for pairwise comparisons of the parasite distribution between the different parasite-fish combinations over time. The caudal-rostral preference of the three parasite strains on the three fish stocks was statistically inferred from these tests (testing the niche partition hypothesis of Gyrodactylus turnbulli and G. bullatarudis for preferences at the caudal and head regions, respectively).
Individual fish after being infected can transition among three discrete host states—fish remains infected (state 1), fish alive with loss of infection (state 2) and fish dead (state 3)—over the observation period. Let ({X_i(t); t ge 0}) be the state of fish i over time. We suppose that ({X_i(t)}) is a time-inhomogeneous Markov chain with transition rate matrix (Q(t)={q_{rs}(t)}) for (r,s=1,2,3). For each (i=1,2, cdots , 157), we have observations (X_i=(X_{i0},X_{i1},cdots ,X_{i9})) at times (t_0=0), (t_1=1), (t_2=3), (cdots) and (t_9=17). The likelihood for the model parameters (theta ={q_{rs}(t)}) is given as
where (L_i(theta |x_i)) is the likelihood contribution for each fish i obtained as product of state transition probabilities
with (p_{x_{ij-1},x_{ij}}(t_{j-1},t_{j})=P{X_i(t_{j})=x_{ij}|X_i(t_{j-1})=x_{ij-1} }.) We assumed that once a fish had lost its infection (state 2) or died (state 3), it could not be reinfected because of the experimental design (move back to state 1) and thus the corresponding rates are 0. Hence, the transition rate matrix Q(t) for the multi-state model with the three discrete host states is given as
where (q_{12}(t)>0) and (q_{13}(t)>0) are the rates at which an infected fish loses its infection and dies at time t respectively. Here, we modelled the rate matrix Q(t) as a piecewise constant function with change points (t_1,t_2,cdots ,t_{8}). For (t{in }[t_{j-1},t_j)), we write (Q(t)=Q_j). The transition probability matrix is
The likelihood function for the model parameters is estimated using a maximum likelihood method, fitted using the msm package in R [60].
We examine how variables such as fish sex, fish size, fish stock and parasite strain may affect the transition rates Q(t). Let (z_i={z_{i1},z_{i2},z_{i3}, z_{i4}}) be the realized values of the covariates (fish sex, fish size, fish stock and parasite strain) for fish i. Then, the transition rate matrix entries (q_{rs}(t)) for (r, s= 1, 2, 3) and (t{in }[t_{j-1},t_j)) were taken as
where ({q_{rsj}^{(0)}}) is baseline intensity; (beta _{rs}) is a parameter vector. The likelihood is then maximized over ({q_{rsj}^{(0)}}) and the regression coefficients (beta _{rs}), for (r=1) and (s=2,3). The hazard ratios (HR) corresponding to each covariate are (exp(beta _{rs})), for (r=1) and (s=2,3). The transition probabilities were estimated from (q_{rs}(t, z_i)) using Eq. 5. Given the four predictors (fish sex, fish size, fish stock and parasite strain) and two possible transitions from state 1 to either state 2 ((q_{12})) or state 3 ((q_{13})) in the proposed multi-state Markov model (defined by Eq. 6), there are (16^2) (or 256) possible variable permutations or models (which includes transitions independent of the underlying covariates).
A systematic variable and model selection was carried out using both Akaike information criterion (AIC) and Bayesian information criterion (BIC) statistics (due to the relative advantages of the two model selection criteria), where all possible variable permutations or models were considered. The AIC statistic assesses the model’s goodness of fit while reducing the complexity of the underlying parameters, whereas the BIC statistics penalise adding more parameters or strongly penalise free parameters compared to the AIC statistic. According to Kuha [61], effective model selection can be achieved by using both AIC and BIC statistics, predominantly to identify models favoured by both criteria, although the study’s methodological design, the main research questions and the belief of a true model and its applicability to the study are crucial factors in determining whether to utilise the AIC or BIC [62]. The best model (among identified parsimonious or highly predictive models) was finally chosen based on a likelihood ratio test (LRT) at a 5% significance level. Detailed results on the variable selection for the multi-state model and its R codes (for reproducibility of results) can be found via the GitHub URL link: github.com/twumasiclement/Spatial-Temporal-Parasite-Dynamics.
Let (T_1) be the time spent in state 1, given that the fish or the process is in state 1 at time 0. Then, the mean sojourn time in state 1 is given as
where
with
and (E(S_j|S_jle t_j-t_{j-1})) is given by Eq. 11 according to Theorem 1. In Eq. 7, the probability that the process leaves in period j, denoted by (text {P(leave in period j)}), is computed such that
with
in accordance to Eq. 10 under Theorem 1.
Let (S_j) be the time spent by infected fish during period j. Suppose that (S_j sim exp left( q_{12}(j,z_i)+q_{13}(j,z_i) right)) with probability density
where (q_{12}(j,z_i)) and (q_{13}(j,z_i)) are the transition rates from state 1 to state 2 and 3, respectively, given the covariates (z_i) for fish i, such that
Then,
and
For the mathematical proof to Theorem 1, see Appendix 2. From Theorem 1, it can be deduced that
where
Also, given the fish or process is in state 1, then the probability of moving to state 2 or 3 next is given as
where
for (s=2,3). We assume that (q_{12}(t,z_i)=q_{12}(15,z_i)) and (q_{13}(t,z_i)=q_{13}(15,z_i)) for (t ge 15).
Fish heatmaps (Fig. 1 and Additional file 1: Fig. S1) depict variations in parasite distribution across eight body regions (caudal fin, lower body, upper body, anal fin, pelvic fin, dorsal fin, pectoral fin and head) over time for each gyrodactylid strain (Gt3, Gt and Gb) on the different fish stocks (OS, LA and UA). Gt3 showed a clear preference for the caudal fin and lower body, with higher mean intensities on OS and LA fish than on the UA stock from day 7 until the end of the infection period. By day 15, all the UA fish had lost the Gt3 infection. Similarly, Gt was more abundant on the tail and lower body until day 13, but switched to a head preference among only OS and LA populations on day 15. In contrast, Gb showed a clear rostral preference from day 7 onwards, a preference strongest in OS(ge)LA(ge)UA fish stocks until the end of the infection period.
When comparing just four body regions of the fish (tail, lower region, upper region and head), the peak time of infection varied spatially across parasite strains and fish stocks (Table 1; Fig. 2). On day 15, higher mean intensities were recorded on the head for both Gt and Gb on OS fish stock. Also for Gb on the same fish stock, a higher number of parasites occurred on the head between days 9 and 17 compared to any other body region or host-parasite combinations (Fig. 2 and Additional file 2: Fig. S2). Parasite distributions varied at the four body regions across the nine host-parasite combinations (Fig. 2) from days 1 to 15 (MKW, (71.25{le }W{le }168.57), (df=32), (p<0.001)), but not on day 17 ((W=38.12), (df=32), (p=0.211)). Only the parasite distribution at the tail and head respectively differed significantly across the nine host-parasite combinations from days 1 to 5 and on day 9 ((ple 0.001)). However, parasite distribution differed significantly among groups on the lower body region on days 7 and 11 (UKW, ({17.12le }H{le }17.74), (df=8), (0.023{le }p{le }0.029)), tail on days 7 and 15 (({19.49le }H{le }24.93), (df=8), (0.002{le }p{le }0.012)) and head on days 7, 11 and 13 ((21.22{le }H{le }47.36), (df=8), (0.001<p{le }0.007)).
From the Bonferroni-Dunn tests, there were significant pairwise differences in parasite distribution at the tail between all Gb groups (Gb-OS, Gb-LA and Gb-UA) and G. turnbulli strains on the fish stocks (with the exception of Gt3 on OS) during day 1 of infection ((0.001< ple)0.016). However, there was no significant difference in parasite distribution of the G. turnbulli strains at the tail across the three fish stocks over time; with the exception of days 3 and 15, between Gt3-OS and Gt-LA groups ((p=0.019)) as well as between Gt3-LA and Gt3-UA groups ((p=0.037)). On days 3 and 5, parasite distribution at the tail was significantly different ((0.001<ple)0.036) between all Gb groups and Gt groups with the exception Gt-OS for day 3 and Gt-UA for day 5. Parasite distribution at the tail on day 7 was significantly different between Gb-UA and Gt groups (Gt-OS and Gt-UA), whilst a significant difference was found between Gb-UA and Gt groups (Gt-LA and Gt-UA). Nevertheless, there was no significant difference between groups of the G. turnbulli strains and G. bullatarudis from day 15 till the end of the infection period. Significant difference in parasite distribution on the lower region only occurred on day 7 between Gt-OS and Gb-UA groups ((p=0.039)) and on day 11 between Gb-UA and Gt-LA groups ((p=0.014)). Nonetheless, parasite distribution on the head was significantly different ((0.001<ple 0.013)) between each of the G. bullatarudis groups (Gb-OS and Gb-UA) and all the G. turnbulli groups on day 1. But from days 3 to 5, significant pairwise difference ((0.001<ple 0.046)) was found between all Gb groups and turnbulli strains for all fish stocks respectively at the head. However, apart from Gb-OS group that still showed significant difference with all G. turnbulli groups on day 7 ((0.001<ple 0.016)), Gb-LA and Gb-UA rather showed significance difference ((0.001<ple 0.037)) with Gt3-OS and Gt3-LA. Nevertheless, Gb on OS showed difference significantly ((0.001<ple 0.013)) on the head with Gt3 on OS and LA stocks as well as Gt on LA population during day 9, whereas two groups of Gb (on OS and LA stocks) had significant difference with only Gt3 on OS fish population during day 11 of the infection period. On day 13, there was significant difference in parasite distribution on the head between Gb and Gt on ornamental fish only.
We used the time-inhomogeneous multi-state Markov model to examine the significant determinants of fish survival (fish sex, fish size, fish stock and parasite strain). The estimated hazard ratios (HR) corresponding to each significant predictor of the fitted model are summarized by Table 2. Figure 3 shows how the baseline transition rates from the infected state (state 1) to uninfected (state 2) and dead (state 3) states changed over the observed time intervals. Figure 4 shows that the fitted multi-state model gives a very good fit to the proportion of fish that will remain in each host infection status from the onset of infection to the end of the study period.
Piecewise-constant plot of estimated baseline transition rates from infected host state to uninfected and dead states at different observed time intervals in the time-inhomogeneous multi-state Markov model
Comparison between observed and expected proportions of fish that will remain in each host infection state from days 1 to 17 after the onset of gyrodactylid infection based on the fitted multi-state Markov model (mean absolute percentage error = 7.85%)
The likelihood of infected fish fighting off their infection was significantly influenced by fish size ((hbox {HR}=0.87), 95% (hbox {C.I}=0.76)-0.99, (p=0.037)), such that larger fish are less likely to clear off their infection. Fish sex, fish stock and parasite strain did influence the likelihood of infected fish dying, but not parasite extinction. Infected male fish were 52% more likely to die compared to female fish ((hbox {HR}=1.52), 95% (hbox {C.I}=1.04)-2.22, (p=0.031)). The risk of death from the gyrodactylid infection among the OS fish ((hbox {HR}=0.24), 95% (hbox {C.I}=0.14)-0.39, (p<0.001)) was 76% less likely compared to UA fish stock. LA fish ((hbox {HR}=0.39), 95% (hbox {C.I}=0.25)-0.61, (p<0.001)) were 61% less likely to die from gyrodactylid infections relative to UA fish. Based on estimated hazard ratios, the rate of fish survival from the gyrodactylid infections was higher among OS stock, followed by LA stock and then UA stock. Fish infected by laboratory strain of G. turnbulli ((hbox {HR}=1.65), 95% (hbox {C.I}=1.03)-2.65, (p=0.037)) were 65% more likely to die compared to the wild strain. The wild G. bullatarudis strain ((hbox {HR}=1.64), 95% (hbox {C.I}=1.02)-2.62, (p=0.039)) was also 64% more likely to kill fish compared to the wild G. turnbulli strain. The estimates of the hazard ratios corresponding to Gt3 and Gb relative to wild G. turnbulli strain suggest that there is no significant difference in the likelihood of fish mortality between Gt3 and Gb strains. We quantified parasite virulence by estimating the rates of both host mortality (Fig. 5) and host recovery (Fig. 6) over time using the fitted multi-state Markov model.
Predicted host mortality rates of parasite strains (Gt3, Gt and Gb) on the fish stocks (OS, LA and UA stocks) over time for both male and female fish respectively
Predicted host recovery rates over time at different fish sizes (11, 14, 17, 20, 23 and 26 mm)
We estimated the mean sojourn time in state 1 (the average amount of time fish can remain infected) and the probability of next transition from the infected state (state 1) to either recovery (state 2) or dead state (state 3) across all significant predictors (fish sex, fish size, fish stock and parasite strain) of the fitted multi-state Markov model. For any strain of gyrodactylid, large ornamental female fish remained infected longer than fish with any other attributes (Table 3). Fish infected with the wild G. turnbulli strain on average remained infected longer than fish infected with Gt3 or wild G. bullatarudis strains before either recovering or dying, irrespective of the fish size, stock and sex. The mean time for fish to remain infected with any parasite strain before fighting off their infection or dying was between 6 and 14 days. An infected fish had a higher probability of dying than recovering from the infection irrespective of the type of gyrodactylid infection, fish stock, sex and size (Table 4). Large male fish were more likely to die than small or medium-sized male or female fish of any size, whereas the chance of host recovery was higher among OS fish stock compared to the Trinidadian fish stocks. The fish infected with wild G. turnbulli strain had a greater probability of fighting off their infections than fish infected with either Gt3 or Gb strain.
In this study, we built on previous studies of the infrapopulation dynamics of three different gyrodactylid strains (two strains of G. turnbulli and one strain of G. bullatarudis) among three different fish stocks (OS, LA and UA stocks) in relation to parasite habitat preference, host survival and parasite virulence (see [45, 48, 49, 56]). We have confirmed for the first time that the microhabitat preferences of the G. turnbulli (laboratory and wild type) and G. bullatarudis strains depend on the type of host and can change over time for the wild G. turnbulli strain (with the help of a multivariate ranked-based distribution-free test and associated post-hoc tests). With an extension to the traditional survival models, we have been able to include host recovery as another absorbing state, and fish sex was identified as a significant factor of host survival compared to previous study of this biological system [45]. We also estimated for the first time, the average duration that fish can remain infectious and the probability that infected fish will either recover or die from each of the three parasite strains across the three guppy populations, sexes and different fish sizes (small, medium and large sizes).
The captive-inbred G. turnbulli strain preferred the tail of three different fish stocks (Ornamental, Lower Aripo River and Upper Aripo River stocks), whereas the wild G. turnbulli initially preferred the tail but then switched to the head. The wild Gyrodactylus bullatarudis consistently showed a rostral preference on all fish. The mean intensity of parasites was higher on OS and LA fish than UA stocks across all body regions over time, probably related to the higher mortality of the UA fish. Lower numbers of parasites on the pectoral, pelvic, dorsal and anal fins compared to the tail, lower body, upper body and head regions might be affected by fish being maintained in isolation or due to difference in the surface area of these body regions. Individual host isolation meant there was no opportunity for host-to-host transmission to occur via the fins (as suggested by [48]). Thus, the parasites might be making a behavioural decision to enhance their fitness in response to the absence of alternative hosts and or reduce competition at small-sized body regions over time. The peak time to infection varied spatially across parasite strains and fish stocks. Such variation likely represents a trade-off between successful parasite exploitation and the host’s localised immune response (reviewed by [34]). Parasite distribution on infected hosts could also be driven by multiple abiotic and biotic factors [50,51,52, 54].
The fitted multi-state model revealed that fish sex, fish stock and parasite strain influenced fish mortality. LA and OS fish stocks survived for longer than UA fish. For this gyrodactylid-fish system, the current study revealed that a longer period of host infection leads to a higher chance of host recovery and a smaller chance of host mortality. The OS guppy population was infectious longer than the Trinidadian fish stocks (LA and UA fish) based on the estimated average duration of infection. However, the OS guppies had a higher chance of host recovery compared to the LA and UA fish stocks, potentially due to superior innate immune defences or immunocompetence towards single-species infections (as revealed by [63, 64]). The LA fish consistently had better parasite resistance than the UA stock across fish sex, parasite strain and different host sizes. Larger fish were infectious over a longer period than small or medium-sized fish, whereas female fish from all three guppy populations experienced a longer duration of infection than male fish. Fish infected by the wild strain of G. turnbulli on average remained infected longer with a higher probability of host recovery among all fish stocks and sexes.
As in the previous study, the laboratory strain of G. turnbulli and wild strain of G. bullatarudis were more likely to cause fish mortality than the wild strain of G. turnbulli, but we found that infected male fish were twice as likely to die relative to female fish. The main reason for this new finding of fish sex as a significant determinant of host mortality is the use of a multi-state model that is able to incorporate host mortality and recovery simultaneously. Other parasite-fish studies have identified fish sex as a significant factor of host mortality [65]. Only fish size significantly influenced the rate of infection loss, namely larger fish acquired more parasites as infections progressed resulting in low parasite extinction compared to smaller fish [53]. Nevertheless, it was found that the chance of host mortality was more likely to occur than host recovery irrespective of host size.
Parasite virulence, described in terms of host mortality and recovery, was significantly time dependent and generally increased towards the end of the infection period. Previously, Gt3 was identified as causing most host deaths, followed by G. bullatarudis and then the wild G. turnbulli, but their respective host mortality rates were not quantified, nor did we previously consider how this changed over time, nor the effect of the different fish stocks [45]. Here, we found no significant difference in host mortality rates between Gt3 and Gb parasite strains over time. Male fish from the three different guppy populations (OS, LA and UA stocks) consistently had a higher rate of host mortality than female fish stocks over time. This could be explained by the fact that the female fish are infectious longer than the male fish as revealed by the estimated mean sojourn time of infection; thus, the female host populations likely develop innate or adaptive host immunity faster than the male fish stocks over time.
The current study could inform the modelling and survival analyses of other biological systems where the entire infection history of an individual (or host) is of interest. Multi-state Markov models provide a robust approach to modelling almost any kind of longitudinal time-to-event data [20]. For multi-state processes that are misclassified or can only be viewed through a noisy marker, hidden Markov models can be implemented [25]. There is more extensive literature on different classes of multi-state Markov models and Markov extension models with specific applications to the modelling of fertility, infectious diseases, competing risks, disability, recurrent events, twin survival and alternating events (reviewed by [15]). However, they have been underused in most parasitological fields.
In the current multi-state Markov model, we could not include spatial information and other relevant information about parasite fecundity, age group (young or old parasite), parasite mortality, parasite mobility and host immune response. A more sophisticated (individual-based) stochastic simulation model would be needed to include these data to further understand the gyrodactylid-fish system. Future studies will examine host-to-host transmission to holistically understand the spread of gyrodactylid parasites and the host-parasite interactions among different populations of fish.
In summary, we identified host-parasite strain-specific microhabitat preferences, discovered determinants of host survival and quantified host-specific parasite virulence based on both host mortality and recovery. The multi-state model was designed so that fish could not be reinfected after infection to match the experimental design, but this could be modified in future studies to include transmission. The multi-state Markov model and the rank-based multivariate Kruskal-Wallis test can be extended and adapted for studying other host-parasite interactions.
The datasets generated and/or analysed during the current study are available from the corresponding author on reasonable request. The R codes used for all statistical analyses have been made publicly available for reproducibility of results via the URL link: https://github.com/twumasiclement/Spatial-Temporal-Parasite-Dynamics.
Akaike information criterion
Bayesian information criterion
Wild Gyrodactylus bullatarudis strain
Wild Gyrodactylus turnbulli strain
Laboratory-bred Gyrodactylus turnbulli strain
Hazard ratio
Individual-based model
Lower Aripo River fish
Likelihood ratio test
Multi-state model
Multivariate Kruskal-Wallis test
Ornamental stock
Population-based model
Univariate Kruskal–Wallis test
Upper Aripo River fish
Milner F, Patton C. A new approach to mathematical modeling of host-parasite systems. Comput Math Appl. 1999;37(2):93–110.
Article  Google Scholar 
Berec L. Techniques of spatially explicit individual-based models: construction, simulation, and mean-field analysis. Ecol Model. 2002;150(1–2):55–81.
Article  Google Scholar 
Adda P, Dimi JL, Iggidir A, Kamgang JC, Sallet G, Tewa JJ. General models of host-parasite systems. Global analysis. Discrete Contin Dyn Syst B. 2007;8(1):1.
Google Scholar 
Edessa GK, Koya PR. Modeling and stability analysis of host-parasite population dynamics. Math Model Appl. 2020;5:28.
Google Scholar 
Gallagher S, Baltimore J. Comparing compartment and agent-based models. In: Eddy WF, editor. Joint statistical meeting, Baltimore; 2017. p. 1–21.
DeAngelis DL. Individual-based models and approaches in ecology: populations, communities and ecosystems. Boca Raton: CRC Press; 2018.
Book  Google Scholar 
Gourbière S, Morand S, Waxman D. Fundamental factors determining the nature of parasite aggregation in hosts. PLoS ONE. 2015;10(2): e0116893.
PubMed  PubMed Central  Article  Google Scholar 
Wilson K, Grenfell BT, Shaw DJ. Analysis of aggregated parasite distributions: a comparison of methods. Funct Ecol. 2006;10(5):592.
Article  Google Scholar 
Caswell H. Matrix population models, vol. 1. Massachusetts: Sinauer Sunderland; 2000.
Google Scholar 
Metz JA, Diekmann O. The dynamics of physiologically structured populations, vol. 68. Berlin: Springer; 2014.
Google Scholar 
Huston M, DeAngelis D, Post W. New computer models unify ecological theory: computer simulations show that many ecological patterns can be explained by interactions among individual organisms. Bioscience. 1988;38(10):682–91.
Article  Google Scholar 
Wallentin G, Neuwirth C. Dynamic hybrid modelling: switching between AB and SD designs of a predator-prey model. Ecol Model. 2017;345:165–75.
Article  Google Scholar 
Abbott RD. Logistic regression in survival analysis. Am J Epidemiol. 1985;121(3):465–71.
CAS  PubMed  Article  Google Scholar 
Efron B. Logistic regression, survival analysis, and the Kaplan-Meier curve. J Am Stat Assoc. 1988;83(402):414–25.
Article  Google Scholar 
Hougaard P. Multi-state models: a review. Lifetime Data Anal. 1999;5(3):239–64.
CAS  PubMed  Article  Google Scholar 
Meira-Machado L, de Uña-Álvarez J, Cadarso-Suárez C, Andersen PK. Multi-state models for the analysis of time-to-event data. Stat Methods Med Res. 2009;18(2):195–222.
PubMed  Article  Google Scholar 
Berry SD, Ngo L, Samelson EJ, Kiel DP. Competing risk of death: an important consideration in studies of older adults. J Am Geriatr Soc. 2010;58(4):783–7.
PubMed  PubMed Central  Article  Google Scholar 
Marshall G, Jones RH. Multi-state models and diabetic retinopathy. Stat Med. 1995;14(18):1975–83.
CAS  PubMed  Article  Google Scholar 
Manzini G, Ettrich TJ, Kremer M, Kornmann M, Henne-Bruns D, Eikema DA, et al. Advantages of a multi-state approach in surgical research: how intermediate events and risk factor profile affect the prognosis of a patient with locally advanced rectal cancer. BMC Med Res Methodol. 2018;18(1):23.
CAS  PubMed  PubMed Central  Article  Google Scholar 
Andersen PK, Borgan Ø. Counting process models for life history data: a review. Preprint series Statistical Research Report. http://urn nb no/URN: NBN: no-23420. 1984;p. 1–104.
Hoem JM, Keiding N, Kulokari H, Natvig B, Barndorff-Nielsen O, Hilden J. The statistical theory of demographic rates: a review of current developments [with discussion and reply]. Scand J Stat. 1976;p. 169–185.
Cox DR, Miller HD. The theory of stochastic processes. Milton Park: Routledge; 2017.
Book  Google Scholar 
Li Y, Cui L, Lin C. Modeling and analysis for multi-state systems with discrete-time Markov regime-switching. Reliabil Eng Syst Saf. 2017;166:41–9.
Article  Google Scholar 
Chakladar S, Liao R, Landau W, Gamalo M, Wang Y. Discrete time multistate model with regime switching for modeling COVID-19 disease progression and clinical outcomes. Stat Biopharm Res. 2022;14(1):52–66.
Article  Google Scholar 
Jackson CH. Multi-state models for panel data: the msm package for R. J Stat Softw. 2011;38(8):1–29.
Article  Google Scholar 
Ferguson N, Datta S, Brock G. msSurv: an R package for nonparametric estimation of multistate models. J Stat Softw. 2012;50:1–24.
Article  Google Scholar 
Courgeau D. Multistate transition models in demography. In: Smelser NJ, Baltes PB, editors. International Encyclopedia of the Social and Behavioral Sciences. Oxford: Pergamon Press; 2001. p. 10210–10214. https://www.sciencedirect.com/science/article/pii/B0080430767021021.
Land KC, Rogers A. Multidimensional mathematical demography: an overview. Multidimension Math Demogr. 1982;p. 1–4.
Courgeau D, Lelièvre E, Lelièvre É. Event history analysis in demography. Oxford: Clarendon Press; 1992.
Google Scholar 
Lelievre E, Bonvalet C, Bry X. Event history analysis of groups. The findings of an on-going research project. Population: an English selection. 1998;p. 11–38.
Lillard LA, Panis CW. aML multilevel multiprocess statistical software, version 2.0. EconWare, Los Angeles, California. 2003. http://www.applied-ml.com.
Bartus T. Multilevel multiprocess modeling with GSEM. Stand Genomic Sci. 2017;17(2):442–61.
Google Scholar 
Shinn A, Pratoomyot J, Bron J, Paladini G, Brooker E, Brooker A. Economic impacts of aquatic parasites on global finfish production. Glob Aquacult Advocate. 2015;2015:58–61.
Google Scholar 
Bakke TA, Cable J, Harris PD. The biology of gyrodactylid monogeneans: the “Russian-Doll Killers’’. Adv Parasitol. 2007;64:161–460.
CAS  PubMed  Article  Google Scholar 
Leung TLF, Bates AE. More rapid and severe disease outbreaks for aquaculture at the tropics: implications for food security. J Appl Ecol. 2013;50(1):215–22.
Article  Google Scholar 
Harris PD, Shinn A, Cable J, Bakke TA. Nominal species of the genus Gyrodactylus von Nordmann 1832 (Monogenea: Gyrodactylidae), with a list of principal host species. Syst Parasitol. 2004;59(1):1–27.
CAS  PubMed  Article  Google Scholar 
Cable J, Archard GA, Mohammed RS, McMullan M, Stephenson JF, Hansen H, et al. Can parasites use predators to spread between primary hosts? Parasitology. 2013;140(9):1138–43.
PubMed  Article  Google Scholar 
Stephenson JF, Van Oosterhout C, Mohammed RS, Cable J. Parasites of Trinidadian guppies: evidence for sex-and age-specific trait-mediated indirect effects of predators. Ecology. 2015;96(2):489–98.
PubMed  Article  Google Scholar 
Griffiths SW, Magurran AE. Sex and schooling behaviour in the Trinidadian guppy. Anim Behav. 1998;56(3):689–93.
CAS  PubMed  Article  Google Scholar 
Croft DP, Edenbrow M, Darden SK, Ramnarine IW, van Oosterhout C, Cable J. Effect of gyrodactylid ectoparasites on host behaviour and social network structure in guppies Poecilia reticulata. Behav Ecol Sociobiol. 2011;65(12):2219–27.
Article  Google Scholar 
Scott ME. Reproductive potential of Gyrodactylus bullatarudis (Monogenea) on guppies (Poecilia reticulata). Parasitology. 1982;85(2):217–36.
Article  Google Scholar 
Malmberg G, et al. The excretory systems and the marginal hooks as a basis for the systematics of Gyrodactylus (Trematoda, Monogenea). Arkiv for Zoologi. 1970;23(1/2):1–235.
Google Scholar 
Llewellyn J. The biology of Isancistrum subulatae n. sp., a monogenean parasitic on the squid, Alloteuthis subulata, at Plymouth. J Mar Biol Assoc UK. 1984;64(2):285–302.
Article  Google Scholar 
Harris PD. Interactions between reproduction and population biology in gyrodactylid monogeneans: a review. B Fr Peche Piscic. 1993;1:47–65.
Article  Google Scholar 
Cable J, van Oosterhout C. The impact of parasites on the life history evolution of guppies (Poecilia reticulata): the effects of host size on parasite virulence. Int J Parasitol. 2007;37(13):1449–58.
CAS  PubMed  Article  Google Scholar 
Rubio-Godoy M, Muñoz-Córdova G, Garduño-Lugo M, Salazar-Ulloa M, Mercado-Vidal G. Microhabitat use, not temperature, regulates intensity of Gyrodactylus cichlidarum long-term infection on farmed tilapia-Are parasites evading competition or immunity? Vet Parasitol. 2012;183(3–4):305–16.
PubMed  Article  Google Scholar 
Ogawa K. A monogenean parasite Gyrodactylus masu sp. n. (Monogenea: Gyrodactylidae) of salmonid fish in Japan. Nippon Suisan Gakkaishi. 1986;52(6):947–950. http://joi.jlc.jst.go.jp/JST.Journalarchive/suisan1932/52.947?from=CrossRef.
Harris PD. Changes in the site specificity of Gyrodactylus turnbulli Harris, 1986 (Monogenea) during infections of individual guppies (Poecilia reticulata Peters, 1859). Can J Zool. 1988;66(12):2854–2857. http://www.nrcresearchpress.com/doi/10.1139/z88-414.
Harris PD, Lyles AM. Infections of Gyrodactylus bullatarudis and Gyrodactylus turnbulli on guppies (Poecilia reticulata) in Trinidad. J Parasitol. 1992;p. 912–914.
Richards G, Chubb J. Host response to initial and challenge infections, following treatment, of Gyrodactylus bullatarudis and G. turnbulli (Monogenea) on the guppy (Poecilia reticulata). Parasitol Res. 1996;82(3):242–7.
CAS  PubMed  Article  Google Scholar 
Hatcher MJ, Dick JT, Dunn AM. How parasites affect interactions between competitors and predators. Ecol Lett. 2006;9(11):1253–71.
PubMed  Article  Google Scholar 
Harvell CD, Mitchell CE, Ward JR, Altizer S, Dobson AP, Ostfeld RS, et al. Climate warming and disease risks for terrestrial and marine biota. Science. 2002;296(5576):2158–62.
CAS  PubMed  Article  Google Scholar 
van Oosterhout C, Potter R, Wright H, Cable J. Gyro-scope: an individual-based computer model to forecast gyrodactylid infections on fish hosts. Int J Parasitol. 2008;38(5):541–8.
PubMed  Article  Google Scholar 
Lu DB, Rudge JW, Wang TP, Donnelly CA, Fang GR, Webster JP. Transmission of Schistosoma japonicum in Marshland and hilly regions of China: Parasite population genetic and sibship structure. PLoS Negl Trop Dis. 2010;4(8).
Mandal FB. Does virulence offer benefit to the parasite ? WebmedCentral Parasitol. 2011;2(10):1–9.
Google Scholar 
Stephenson JF, Young KA, Fox J, Jokela J, Cable J, Perkins SE. Host heterogeneity affects both parasite transmission to and fitness on subsequent hosts. Phil Trans R Soc B Biol Sci. 2017;372(1719):1–10.
Article  Google Scholar 
R Core Team, et al.. Hurley C, editor. R: a language and environment for statistical computing. Vienna, Austria: R Core Team; 2019. http://www.R-project.org/.
Kimball S, Mattis P. Gnu image manipulation program (gimp). GIMP. 2012;6:1–653. https://www.gimp.org/.
He F, Mazumdar S, Tang G, Bhatia T, Anderson SJ, Dew MA, et al. Non-parametric MANOVA approaches for non-normal multivariate outcomes with missing values. Commun Stat Theory Methods. 2017;46(14):7188–200.
PubMed  PubMed Central  Article  Google Scholar 
Jackson CH, Sharples LD, Thompson SG, Duffy SW, Couto E. Multistate Markov models for disease progression with classification error. J R Stat Soc Ser D (The Statistician). 2003;52(2):193–209.
Article  Google Scholar 
Kuha J. AIC and BIC: comparisons of assumptions and performance. Sociol Methods Res. 2004;33(2):188–229.
Article  Google Scholar 
Vrieze SI. Model selection and psychological theory: a discussion of the differences between the Akaike information criterion (AIC) and the Bayesian information criterion (BIC). Psychol Methods. 2012;17(2):228.
PubMed  PubMed Central  Article  Google Scholar 
Magurran AE, Seghers BH. Population differences in predator recognition and attack cone avoidance in the guppy Poecilia reticulata. Anim Behav. 1990;40(3):443–52.
Article  Google Scholar 
Smallbone W, Ellison A, Poulton S, van Oosterhout C, Cable J. Depletion of MHC supertype during domestication can compromise immunocompetence. Mol Ecol. 2021;30(3):736–46.
CAS  PubMed  Article  Google Scholar 
Wileman D, Sangster G, Breen M, Ulmestrand M, Soldal A, Harris R. Roundfish and Nephrops survival after escape from commercial fishing gear. EC Contract No: FAIR-CT95-0753 Final Report. 1999 11;p. 1–241.
Download references
Not applicable.
This study was funded by a Cardiff University Vice Chancellor’s international scholarship scheme for research excellence to C.T.
School of Mathematics, Cardiff University, Abacws, CF24 4AG, Cardiff, UK
Clement Twumasi & Owen Jones
School of Biosciences, Cardiff University, Sir Martin Evans Building, CF10 3AX, Cardiff, UK
Clement Twumasi & Joanne Cable
Department of Statistics, Oxford University, 29 St Giles’, OX1 3LB, Oxford, UK
Clement Twumasi
You can also search for this author in PubMed Google Scholar
You can also search for this author in PubMed Google Scholar
You can also search for this author in PubMed Google Scholar
C.T. and O.J. developed the mathematical methodologies. J.C. provided the empirical data. C.T. performed all statistical analyses and drafted the article. All authors conceived and designed the study, read and approved the final manuscript.
Correspondence to Clement Twumasi.
Not applicable.
Not applicable.
All authors declare that there are no conflicts of interest.
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Detailed visualisation of fish heatmaps over eight body regions of fish across parasite strains and fish stocks over time (from day 1 to 17).
Grouped barcharts showing variations in mean intensities at four main body regions of fish across parasite strains and fish stocks over surviving fish and across time (from day 1 to 17).
For simplicity, let (S_j sim exp left( lambda right)) with probability density (f(S_j)= lambda mathrm{e}^{-lambda S_j}, quad S_j>0) and (E(S_j)=frac{1}{lambda }), where (lambda =q_{12}(j,z_i)+q_{13}(j,z_i)). Suppose (alpha =t_j-t_{j-1}), then
Considering the integral I and using integration by parts,
where (u=S_j), (u^{prime }=1), (v^{prime }=mathrm{e}^{-lambda S_j}) and (v=-frac{1}{lambda }mathrm{e}^{-lambda S_j}).
Hence,
Substituting for (lambda) and (alpha) gives the required Eq. 9 such that
Also, the required Eq. 10 can be obtained such that
(square)
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated in a credit line to the data.
Reprints and Permissions
Twumasi, C., Jones, O. & Cable, J. Spatial and temporal parasite dynamics: microhabitat preferences and infection progression of two co-infecting gyrodactylids. Parasites Vectors 15, 336 (2022). https://doi.org/10.1186/s13071-022-05471-9
Download citation
Received:
Accepted:
Published:
DOI: https://doi.org/10.1186/s13071-022-05471-9
Anyone you share the following link with will be able to read this content:
Sorry, a shareable link is not currently available for this article.

Provided by the Springer Nature SharedIt content-sharing initiative
Advertisement
ISSN: 1756-3305
By using this website, you agree to our Terms and Conditions, California Privacy Statement, Privacy statement and Cookies policy. Manage cookies/Do not sell my data we use in the preference centre.
© 2022 BioMed Central Ltd unless otherwise stated. Part of Springer Nature.

source

Entrepreneur Teague Egan EnergyX Company – stupidDOPE.com


The global battle for renewable energy resources is here, particularly battery metals and materials. The American startup, EnergyX, is leading the charge in revolutionizing the energy storage market with the development of sustainable lithium refining and battery technology.
Behind this clean energy powerhouse is an unlikely figure, 34 year-old founder and CEO, Teague Egan. The serial entrepreneur, quickly becoming an energy mogul, has often been compared to Elon Musk. He was inspired to build the lithium company after a life-changing trip to South America in 2018. On the advice of a friend, Egan decided to backpack the picturesque Salar de Uyuni in Bolivia, the world’s largest salt flat at over 4,000 square miles.
While touring the salt pan, Egan learned that this area making up part of the “Lithium Triangle” was also the world’s largest lithium reserve. As an early investor in Tesla, he had strong conviction that the world was shifting to electric vehicles, which meant massive quantities of lithium. What he didn’t know was that the technology to drastically increase the supply of this key battery metal was not yet commercially available.

Who is Teague Egan?
A South Florida native, Teague Egan found early success in the entertainment industry while attending University of Southern California. Earning his undergraduate degree in business entrepreneurship, he started a record label out of his dorm room, and his first album release quickly landed at #1 on iTunes.
The son of early internet investor and entrepreneur Michael Egan, Teague attributes his entrepreneurial spirit to early childhood conversations with his father. “When I was 12 years old, dinner table conversations were about balance sheets, EBITDA, and complex problem solving,” Egan explains “as well as the trials and tribulations of entrepreneurship.”
Fast forward to 28 years old, Michael told Teague to write down his 5 passions and the 5 industries he thought would be the biggest in the next 20-40 years. Now known as the 5&5 Exercise, the energy transition was something that overlapped on Egan’s list. With a lot of passion, precise timing, and a little luck, Egan found himself standing in the Saudi Arabia of lithium.
 
A Moment of Inspiration
Ultimately determining that the climate crisis was where Egan wanted to focus his time and energy, that trip to Bolivia was the juncture of vision and inspiration that drove Egan to leave entertainment and commit everything to solving the lithium supply chain issues. Within 4 years, EnergyX has catapulted into one of the leaders in Direct Lithium Extraction (DLE) and lithium refinery, boasting partnerships with some of the largest lithium producers in the world.
The company won an international tender to deploy its pilot to Bolivia, and successfully operated the first ever DLE technology pilot plant in the Lithium Triangle with 94% recovery rate, only to be disqualified for turning in its final report 10 minutes late. There were rumors of external geopolitical factors involving Russia and China, the other competing entities, were at play.
Now with over $450M in funding commitments, and already featuring an all-star team of PhD’s, chemical engineers, and clean energy veterans, EnergyX most recently announced an interesting addition to their advisory board, A-List super-firm Endeavor, whose fearless leader is none other than Ari Emanuel.

Full Circle with Ari Emanuel
During Egan’s first career in Hollywood, his record label represented breakout musician Sammy Adams who performed at shows across the nation to sell out crowds. Coincidentally, Egan’s artists were represented by Endeavor’s William Morris Agency (WME) whose talent roster also includes the likes of musicians Usher, Bruno Mars, and Foo Fighters, actors Denzel Washington and Jessica Alba, sports stars such as Tom Brady, and a host of other talents.
The Endeavor Group has also committed to ‘Endeavor Impact’, a company-wide commitment to equity, democracy, inclusion, and sustainability that is a core element of this advisory. “It was literally 10 years after the record label, and I got a call that said WME on my caller ID,” says Egan. “When I picked up, it was Ari Emanuel himself asking about EnergyX and our featured cover story on the front page of New York Times.”
With a mission focused on sustainability, Endeavor’s team will help drive awareness and strategic conversation around the future of energy. It’s no doubt that the firm’s cultural footprints in the world of entertainment, sports, and media could be key to bridging EnergyX’s vision and technology to help combat the climate crisis and quicken the world’s transition to renewable energy.
A New Approach

While lithium is the critical component in electric vehicle batteries, the extraction of this metal has caused some serious environmental and social concerns over the years. Egan hopes to break this narrative by placing sustainability at the heart of technological innovations in lithium extraction. Elon Musk recently commented in Tesla’s Q2 earnings that “lithium refining is like having a license to print money.”
Over the last year, the metal’s price has skyrocketed, going from $10,000 to $80,000 per ton influenced by demand from large car companies such as GM, Ford and Tesla, in addition to new government regulations pushing for Net-Zero and Low Carbon initiatives.
From entertainment breakout to energy titan, Egan and team are now revolutionizing the renewable energy industry. Through an unlikely partnership between Egan and Emanuel, they hope that Endeavor Impact’s mission, A-list celebrities, and the savvy marketing and influence of Endeavor can help to spread the important message of fighting climate change.
Whether this duo can crack the code to sustainable lithium extraction is at the center of crucial questions on climate, energy, and the future of technology. Learn more about Teague Egan, and the future of lithium at EnergyX.
I am a Singer, Songwriter, College Graduate with BBA & Co-Founder/Editor-In-Chief of stupidDOPE.com. I was born and raised in South Carolina, currently live in Florida.
Subscribe now to our newsletter

Input your search keywords and press Enter.

source

Dodgers vs. Cardinals live updates: Albert Pujols hits No. 700 – Los Angeles Times

Craig Kimbrel told he won’t be Dodgers’ closer before Albert Pujols’ historic game
By Steve Henson
Craig Kimbrel is no longer the Dodgers’ closer.
There. Dave Roberts said so Friday. Finally.
The words came haltingly, with thoughtful pauses, but they came.
“I just think that right now, um, I’m going to keep Craig down tonight,” Roberts said. “And my expectation is, yeah, I talked to him today. Right now the plan is to change roles and get him into a position to pitch in different innings in different situations.
“He was very open to doing whatever is best for the ballclub. So I feel good about it and we’ll see where that takes us.”
Dodgers
Mookie Betts’ pinch-hit single lifted the Dodgers to a 3-2 walk-off win over the Arizona Diamondbacks on Thursday at Dodger Stadium.

Roberts comments came ahead of the Dodgers’ 11-0 loss to the St. Louis Cardinals at Dodger Stadium highlighted by Albert Pujols joining Barry Bonds, Hank Aaron and Babe Ruth in the 700 home run club. The former Dodgers and Angels slugger hit a two-run homer off Andrew Heaney in the third inning and a three-run shot off Phil Bickford in the fourth to the delight of a crowd of 50,041.
When the sixth inning rolled around, Roberts contradicted his pregame statement by calling on Kimbrel with the Dodgers trailing 8-0. Is the new role of a reliever with 394 career saves mop-up man? It’s hard to know how best to use him. Kimbrel did retire Pujols on a groundout to shortstop.
Kimbrel gave up a home run in the ninth inning Thursday that put the Arizona Diamondbacks ahead by a run. The Dodgers won in the bottom of the inning on Mookie Betts’ walk-off single, which gave Kimbrel the win. That clearly was the last straw for Roberts.
Kimbrel, 34, has blown five of 27 save opportunities and has a 4.07 ERA in 58 appearances this season. Only 12.1% of his pitches swung at are missed, about league average. Yet he is seventh all-time in career saves and led the league in saves four consecutive years ending in 2014.
Perhaps that’s why Roberts and the Dodgers front office stuck with him so long even though his effectiveness had obviously diminished from his heyday.
Roberts said he appreciated the way Kimbrel took the news that he was no longer the closer.
“My job is to find the best lane for him and he has no worries of changing the routine from something he’s always done,” Roberts said. “He looks at it as a challenge. That’s the way we would expect it and it speaks a lot to him as a team player.”
Roberts said no one reliever will assume the closer role for now. Evan Phillips (6-3, 1.24 ERA), Brusdar Graterol (2-3, 2.96) and Alex Vesia (4-0, 2.26) are the obvious candidates.
Obituaries
Maury Wills, who revolutionized baseball with his ability to steal bases and was the MVP of the National League in 1962 for the Dodgers, has died.

“No, for us it’s finding, it’s treating him like we treat all of our guys, putting them in the best position to get outs,” Roberts said. “That’s kind of how I’m going to approach every inning for the ‘pen.”
Phillips, 28, has been the undisputed top reliever all season, giving up 29 hits in 58 innings while striking out 68. He’s in line for closer duty, although his value has been putting out fires regardless of the inning.
Also, he has only three career saves, two this season. Phillips said he wouldn’t treat the ninth inning differently than any other.
“I think the best thing you can do as the guy coming out to finish the game is to provide consistency throughout the season,” he said.
If a closer-by-committee approach fails, the Dodgers face a situation where a new closer must emerge with the postseason around the corner. Phillips didn’t sound like he would be fazed.
“I don’t really value that role any higher than what I do now,” he said. “I think the [Dodgers] really instilled in us that when the phone rings you go out there and get your outs. So regardless of the situation, and it might be the ninth inning or a tight ballgame, however you want to phrase it, my challenge is that when the phone rings, I do my job.”
Kimbrel likely will be used almost exclusively against right-handed hitters, who are batting .208 with a .603 on-base-plus-slugging percentage against him. Left-handers are hitting .266 with a .786 OPS.
Silencing the Dodgers on Friday was longtime nemesis Jose Quintana, who pitched six scoreless innings, reducing his ERA in 12 career games against the Dodgers to 1.80. The left-hander struggled to a 6.75 ERA in 24 appearances with the Angels in 2021 but has posted a 3.04 ERA in 30 starts with the Cardinals and Pittsburgh Pirates this season.
Roberts said injured starter Tony Gonsolin had normal soreness “which is a good thing” a day after he pitched two innings in a simulated game at Dodger Stadium. Roberts said Gonsolin is scheduled to pitch two innings in a rehab game at triple-A Oklahoma City.
“
Another home run for the Cardinals
It’s the eighth inning and Hanser Alberto is pitching for the Dodgers. It’s officially become a don’t-waste-the-bullpen game for the Dodgers.
Alec Burleson took advantage of that fact by hitting his first career home run — a solo shot to right field — to give the Cardinals an 11-0 lead. It was the fifth home run of the game for the Cardinals.
Alec Burleson collects his first MLB homer! pic.twitter.com/TKn5T3rmsv
The Dodgers are on pace for their worst shutout loss since 2018.
⚾ End of eighth: Cardinals 11, Dodgers 0
Cardinals homer again, take 10-0 lead in the seventh
Albert Pujols isn’t the only player on the Cardinals hitting home runs tonight.
Juan Yepez hit his 12th home run of the season off Dodgers reliver Caleb Ferguson, slinging a 95-mph fastball just inside the left-field foul pole to help the Cardinals take a 10-0 lead in the seventh inning. It was the fourth home run hit by the Cardinals.
After Yepez’s solo shot made it 9-0, Lars Nootbar drove in another run two batters later on a grounder to center field.
The Dodgers certainly don’t look like a 100-win team tonight. With all the woes the bullpen has been experiencing — including the situation involving Craig Kimbrel — is it cause for concern or an anomaly on dominance street?
⚾ Middle of the seventh: Cardinals 10, Dodgers 0
Cardinals continue to hammer Dodgers pitching
After Albert Pujols punched the Cardinals to a 5-0 lead with his 699th and 700th career home runs, St. Louis continued to pour it on against the Dodgers.
The Cardinals scored three more runs off Phil Bickford in the sixth inning. Dylan Carlson doubled in a run before Lars Nootbaar hit a two-run homer to right field to give the Cardinals an 8-0 lead.
Albert Pujols hits home run No. 700
Cardinals slugger Albert Pujols hit his 700th career home run in the fourth inning off Dodgers reliever Phil Bickford, giving St. Louis a 5-0 lead.
The three-run shot was another no-doubter, landing in the left-field pavilion. Dodger Stadium erupted in cheers with players on the Cardinals congratulating Pujols and Dodgers players clapping in appreciation.
7️⃣0️⃣0️⃣ #Pujols700 pic.twitter.com/XyddSFPpcQ
Pujols then came out of the dugout and acknowledged the crowd.
Read more >>>
Albert Pujols hits home run No. 699
Albert Pujols hit his 699th career home run in his second at-bat against the Dodgers in the third inning Friday night, clobbering a pitch from Andrew Heaney into the pavilion in left-center field at Dodger Stadium.
No doubt about number 699 for #AlbertPujols! pic.twitter.com/2Ns3ueuzcn
It was a no-doubter, Pujols extracting everything on the pitch for a two-run blast deep into the pavilion seats, giving the Cardinals a 2-0 lead.
Pujols is one home run away from joining Barry Bonds (762), Hank Aaron (755) and Babe Ruth (714) as the only players in major league history to hit 700 home runs.
Albert Pujols hits No. 699 and Dodger Stadium cheers. pic.twitter.com/I7chdj5mWd
⚾ Middle of third: Cardinals 2, Dodgers 0
Albert Pujols strikes out in first at-bat vs. Dodgers
By Austin Knoblauch
Dodgers starter Andrew Heaney struck out St. Louis Cardinals slugger Albert Pujols with a 95-mph fastball during his at-bat in the first inning.
Pujols, sitting on 698 career home runs, received a standing ovation from many of the fans in attendance at Dodger Stadium, where Pujols spent most of his 2021 season.
Pujols is looking to become only the fourth player in major league history to hit 700 home runs. He said will retire at the end of the season.
⚾ Middle of the first: Cardinals 0, Dodgers 0
Craig Kimbrel ousted as Dodgers closer
By Steve Henson
Craig Kimbrel is no longer the Dodgers closer.
There. Dave Roberts said so Friday. Finally.
The words came haltingly, with thoughtful pauses, but they came.
“I just think that right now, um, I’m going to keep Craig down tonight,” Roberts said. “And my expectation is, yeah, I talked to him today. Right now the plan is to change roles and get him into a position to pitch in different innings in different situations.
“He was very open to doing whatever is best for the ballclub. So I feel good about it and we’ll see where that takes us.”
Kimbrel gave up a home run in the ninth inning Thursday that put the Arizona Diamondbacks ahead by a run. The Dodgers won in the bottom of the inning on Mookie Betts’ walk-off single, which gave Kimbrel the win.
Dodgers
Mookie Betts’ pinch-hit single lifted the Dodgers to a 3-2 walk-off win over the Arizona Diamondbacks on Thursday at Dodger Stadium.

But Kimbrel, 34, has blown five of 27 save opportunities and has a 4.14 ERA in 57 appearances this season. He is seventh all-time with 394 saves in 13 seasons and led the league in saves four consecutive years ending in 2014.
Perhaps that’s why Roberts and the Dodgers front office stuck with him so long even though his effectiveness had obviously diminished from his heyday.
Roberts said he appreciated the way Kimbrel took the news that he was no longer the closer.
“He’s such a professional and knows that there is more there, more consistency, just the openness to do what’s best,” he said. “My job is to find the best lane for him and he has no worries of changing the routine from something he’s always done. He looks at it as a challenge. That’s the way we would expect it and it speaks a lot to him as a team player.”
Roberts said no one reliever will assume the closer role for now. Evan Phillips, Brusdar Graterol and Alex Vesia are the most obvious candidates.
“No, for us it’s finding, it’s treating him like we treat all of our guys, putting them in the best position to get outs,” Roberts said. “That’s kind of how I’m going to approach every inning for the ‘pen.”

Rejuvenated Albert Pujols brings his chase for 700 home runs to Dodger Stadium
By Mike DiGiovanna
SAN DIEGO — When the bat of Albert Pujols caught fire in early August, and it appeared the St. Louis Cardinals slugger would make an improbable 11th-hour run to 700 homers, Mark McGwire began tracking his long-ago teammate through the MLB app and tuning into as many Pujols plate appearances as he could.
“His at-bats, he looks like he’s 25 again,” McGwire said of the 42-year-old Pujols, who was a Cardinals rookie when McGwire hit the last of his 583 career homers in a St. Louis uniform in 2001. “This guy is a born hitter.”
But it wasn’t until McGwire returned to Busch Stadium for Matt Holliday’s Cardinals Hall-of-Fame induction ceremony on Aug. 27 that he truly grasped — literally, not figuratively — how Pujols, in his 22nd and final big league season, has recaptured the thunder in his swing.
Read more >>>
ICYMI: Mookie Betts lifts Dodgers with walk-off win over Diamondbacks
By Steve Henson
Four hours before the Dodgers defeated the Arizona Diamondbacks 3-2 Thursday on a walk-off hit by Mookie Betts, the eyes of everyone in blue were riveted on the Dodger Stadium pitcher’s mound.
Manager Dave Roberts and bench coach Bob Geren stood near the third-base line, arms folded. President of Baseball Operations Andrew Friedman watched while engaged in conversation with Betts. A good half-dozen Dodgers pitchers leaned against the dugout railing, observing every pitch.
Nobody needed to state the obvious: What they were witnessing could seriously impact their postseason fortunes.
Tony Gonsolin, on the injured list since Aug. 29 with a forearm strain, was pitching to teammates Gavin Lux, Chris Taylor, Hanser Alberto and Miguel Vargas. The All-Star right-hander held nothing back, throwing his four-seam fastball, slider, curveball and split-finger fastball.
Read more >>>
How to watch and stream the Dodgers this season
Here’s a look at the Dodgers broadcast and streaming schedule for the remainder of the 2022 regular season:

Live Feed
The Dodgers (104-47) lost to the St. Louis Cardinals (89-63) at Dodger Stadium on Friday. Here’s a recap of the game.

Subscribe for unlimited access
Follow Us

source

Interview: Egzod Talks about His New Single “The Way Out” – CelebMix




Melodic bass producer Egzod releases his brand-new single, “The Way Out,” featuring Bertie Scott, via Spinnin’ Records. “The Way Out” delivers a cinematic journey, exploring the emotions of “something slipping away.”
Born in Portugal, raised in France, and now based in the U.S., Egzod has amassed more than 200 million streams and millions of international fans, including over two million monthly listeners on Spotify.
Egzod, which means “Exodus” in French phonetics, is the narrative of a mass departure of people to another world where technology and nature are in perfect balance. Through his music, blending sounds and instruments from around the globe, Egzod works to empower and unify people.
Appearing soon at Firefly Music Fest and EDC Orlando, Egzod’s new tracks will provide a “way out” towards exciting musical fantasies.
CelebMix spoke with Egzod to discover more about the person behind the music, how he got started in music and his definition of tone.

What inspired your new single, “The Way Out?”
I have been finding a lot of inspiration in Hardwave in the past years as it is very cinematic but also combines both of my favorite influences: the melodic content from Trance and the heaviness of Bass music. Bertie Scott sent over some toplines and “The Way Out” fitted perfectly on a track I was working on, both lyrically and melodically.
Where and when did you connect with Bertie Scott?
I connected with Bertie Scott via Instagram last year. We’ve been going back and forth on different music ideas ever since. We recently finished another track together actually.
How did you get started in music?
I first got into making music through the art of DJing at the age of 13. I’d spend time making mashups and mixes up until the end of high school. Starting college, I switched to music production because I wanted to start making my own records.
Where are you from?
I was born in Portugal, close to Porto, and I grew up in the South of France. I’m currently based in Florida.
Did your hometown impact your music?
It did quite a lot. Electronic music has always been big in France – from the French Touch scene in the ‘90s/early 2000s with artists like Daft Punk, Justice, David Guetta, Martin Solveig, Breakbot – and now more recently in bass and house music in the past decade with artists like DJ Snake, Dirtyphonics, Habstrakt, Malaa, Tchami, CloZee. However, I’d say what influenced me more than my hometown, is the internet. I think that a lot of subgenres like Future Bass, Trap, Hardwave, and Dubstep exploded thanks to online forums and curators/labels on Youtube and other platforms in the past decade.
Did your sound evolve naturally, or did you deliberately push it in a certain direction?
I’d say that I pushed it to a certain direction, being bass music blended with some orchestral elements. Then from there, my sound evolved naturally. I think it’s a never-ending process which makes music production always exciting.
How do you define tone? And has your tone changed over time?
Lately, it has shifted from an orchestral/eastern tone to a more futuristic/hybrid style I would say. Although, I tend to always keep it melodic. As mentioned previously, I think it’s a never-ending process and an artist’s style or tone evolves forever, although the essence is kept.
What inspires your writing? Do you draw inspiration from poems, music, or other media?
Definitely a lot of fantasy and sci-fi movies (ex: LOTR, Ghost In The Shell, Matrix)  – but also digital art and 3D animation (ex: Beeple). I listen to all music genres on a daily basis which helps me find inspiration as well.
What can you share about your writing process?
I love resampling my own work that has never been released. Some of my biggest releases have come out of old projects that I would recycle or resample. Tip for the producers out there: keep ALL your works in progress stored – you’d be surprised what could happen with them years down the road.
Which artists in your opinion are killing it right now?
I’m maybe biased because I’ve collabed with most of them but I’m genuinely a fan of their work. Here is the list in no particular order: if found, Afinity, Slippy, Arcando, Memba, Maazel, Nurko, Abandoned, Madzi, and Sublab.
How do you define success?
Success is the accomplishment of a purpose. Part of society has put a lot of focus on success being exclusively related to numbers and the achievement of goals. I’d say it’s not exclusive to that and goes beyond those. Goals are a point of view to achieve – a purpose is the reason why we want to achieve these goals. I think finding the reason is what brings success.
What can your fans look forward to over the next six months? Music videos? Live gigs?
I’m very stoked to be playing at Firefly for the first time and EDC Orlando this year. I’m excited to be making my debut on Spinnin’ Records as well with “The Way Out” which will be released with its lyric video. Expect lots of new originals and remixes following this release!
Follow Egzod Instagram | Twitter | Facebook | Spotify
Randy Radic is a Left Coast author and writer. Author of numerous true crime books written under the pen-name of John Lee Brook. Former music contributor at Huff Post.
Read More
View Gallery
Read More
Read More
View Gallery
Read More
Read More
Read More
Read More
Read More
2022 CelebMix
Sağlık Personeli

source

‹ Previous Next ›