Coalescent inference for infectious disease: meta-analysis of hepatitis C

Estimating the fitness cost and benefit of antimicrobial resistance from pathogen genomic data

Increasing levels of antibiotic resistance in many bacterial pathogen populations are a major threat to public health. Resistance to an antibiotic provides a fitness benefit when the bacteria are exposed to this antibiotic, but resistance also often comes at a cost to the resistant pathogen relative to susceptible counterparts. We lack a good understanding of these benefits and costs of resistance for many bacterial pathogens and antibiotics, but estimating them could lead to better use of antibiotics in a way that reduces or prevents the spread of resistance. Here, we propose a new model for the joint epidemiology of susceptible and resistant variants, which includes explicit parameters for the cost and benefit of resistance. We show how Bayesian inference can be performed under this model using phylogenetic data from susceptible and resistant lineages and that by combining data from both we are able to disentangle and estimate the resistance cost and benefit parameters separately. We applied our inferential methodology to several simulated datasets to demonstrate good scalability and accuracy. We analysed a dataset of Neisseria gonorrhoeae genomes collected between 2000 and 2013 in the USA. We found that two unrelated lineages resistant to fluoroquinolones shared similar epidemic dynamics and resistance parameters. Fluoroquinolones were abandoned for the treatment of gonorrhoea due to increasing levels of resistance, but our results suggest that they could be used to treat a minority of around 10% of cases without causing resistance to grow again.

Model design for nonparametric phylodynamic inference and applications to pathogen surveillance

Inference of effective population size from genomic data can provide unique information about demographic history and, when applied to pathogen genetic data, can also provide insights into epidemiological dynamics. The combination of nonparametric models for population dynamics with molecular clock models which relate genetic data to time has enabled phylodynamic inference based on large sets of time-stamped genetic sequence data. The methodology for nonparametric inference of effective population size is well-developed in the Bayesian setting, but here we develop a frequentist approach based on nonparametric latent process models of population size dynamics. We appeal to statistical principles based on out-of-sample prediction accuracy in order to optimize parameters that control shape and smoothness of the population size over time. Our methodology is implemented in a new R package entitled mlesky. We demonstrate the flexibility and speed of this approach in a series of simulation experiments and apply the methodology to a dataset of HIV-1 in the USA. We also estimate the impact of non-pharmaceutical interventions for COVID-19 in England using thousands of SARS-CoV-2 sequences. By incorporating a measure of the strength of these interventions over time within the phylodynamic model, we estimate the impact of the first national lockdown in the UK on the epidemic reproduction number.

Genomic epidemiology of Salmonella Typhi in Central Division, Fiji, 2012 to 2016

BACKGROUND

Typhoid fever is endemic in some Pacific Island Countries including Fiji and Samoa yet genomic surveillance is not routine in such settings. Previous studies suggested imports of the global H58 clade of Salmonella enterica var Typhi (Salmonella Typhi) contribute to disease in these countries which, given the MDR potential of H58, does not auger well for treatment. The objective of the study was to define the genomic epidemiology of Salmonella Typhi in Fiji.

METHODS

Genomic sequencing approaches were implemented to study the distribution of 255 Salmonella Typhi isolates from the Central Division of Fiji. We augmented epidemiological surveillance and Bayesian phylogenomic approaches with a multi-year typhoid case-control study to define geospatial patterns among typhoid cases.

FINDINGS

Genomic analyses showed Salmonella Typhi from Fiji resolved into 2 non-H58 genotypes with isolates from the two dominant ethnic groups, the Indigenous (iTaukei) and non-iTaukei genetically indistinguishable. Low rates of international importation of clones was observed and overall, there were very low levels an antibiotic resistance within the endemic Fijian typhoid genotypes. Genomic epidemiological investigations were able to identify previously unlinked case clusters. Bayesian phylodynamic analyses suggested that genomic variation within the larger endemic Salmonella Typhi genotype expanded at discreet times, then contracted.

INTERPRETATION

Cyclones and flooding drove 'waves' of typhoid outbreaks in Fiji which, through population aggregation, poor sanitation and water safety, and then mobility of the population, spread clones more widely. Minimal international importations of new typhoid clones suggest that targeted local intervention strategies may be useful in controlling endemic typhoid infection. These findings add to our understanding of typhoid transmission networks in an endemic island country with broad implications, particularly across Pacific Island Countries.

FUNDING

This work was supported by the Coalition Against Typhoid through the Bill and Melinda Gates Foundation [grant number OPP1017518], the Victorian Government, the National Health and Medical Research Council Australia, the Australian Research Council, and the Fiji Ministry of Health and Medical Services.

Epidemiological inference from pathogen genomes: A review of phylodynamic models and applications

Phylodynamics requires an interdisciplinary understanding of phylogenetics, epidemiology, and statistical inference. It has also experienced more intense application than ever before amid the SARS-CoV-2 pandemic. In light of this, we present a review of phylodynamic models beginning with foundational models and assumptions. Our target audience is public health researchers, epidemiologists, and biologists seeking a working knowledge of the links between epidemiology, evolutionary models, and resulting epidemiological inference. We discuss the assumptions linking evolutionary models of pathogen population size to epidemiological models of the infected population size. We then describe statistical inference for phylodynamic models and list how output parameters can be rearranged for epidemiological interpretation. We go on to cover more sophisticated models and finish by highlighting future directions.

Computing R 0 of dynamic models by a definition-based method

Objectives

Computing the basic reproduction number (R 0) in deterministic dynamical models is a hot topic and is frequently demanded by researchers in public health. The next-generation methods (NGM) are widely used for such computation, however, the results of NGM are usually not to be the true R 0 but only a threshold quantity with little interpretation. In this paper, a definition-based method (DBM) is proposed to solve such a problem.

Methods

Start with the definition of R 0, consider different states that one infected individual may develop into, and take expectations. A comparison with NGM has proceeded. Numerical verification is performed using parameters fitted by data of COVID-19 in Hunan Province.

Results

DBM and NGM give identical expressions for single-host models with single-group and interactive R ij of single-host models with multi-groups, while difference arises for models partitioned into subgroups. Numerical verification showed the consistencies and differences between DBM and NGM, which supports the conclusion that R 0 derived by DBM with true epidemiological interpretations are better.

Conclusions

DBM is more suitable for single-host models, especially for models partitioned into subgroups. However, for multi-host dynamic models where the true R 0 is failed to define, we may turn to the NGM for the threshold R 0.

Statistical Challenges in Tracking the Evolution of SARS-CoV-2

Genomic surveillance of SARS-CoV-2 has been instrumental in tracking the spread and evolution of the virus during the pandemic. The availability of SARS-CoV-2 molecular sequences isolated from infected individuals, coupled with phylodynamic methods, have provided insights into the origin of the virus, its evolutionary rate, the timing of introductions, the patterns of transmission, and the rise of novel variants that have spread through populations. Despite enormous global efforts of governments, laboratories, and researchers to collect and sequence molecular data, many challenges remain in analyzing and interpreting the data collected. Here, we describe the models and methods currently used to monitor the spread of SARS-CoV-2, discuss long-standing and new statistical challenges, and propose a method for tracking the rise of novel variants during the epidemic.

Markov genealogy processes

We construct a family of genealogy-valued Markov processes that are induced by a continuous-time Markov population process. We derive exact expressions for the likelihood of a given genealogy conditional on the history of the underlying population process. These lead to a nonlinear filtering equation which can be used to design efficient Monte Carlo inference algorithms. We demonstrate these calculations with several examples. Existing full-information approaches for phylodynamic inference are special cases of the theory.

Employing phylogenetic tree shape statistics to resolve the underlying host population structure

Background

Host population structure is a key determinant of pathogen and infectious disease transmission patterns. Pathogen phylogenetic trees are useful tools to reveal the population structure underlying an epidemic. Determining whether a population is structured or not is useful in informing the type of phylogenetic methods to be used in a given study. We employ tree statistics derived from phylogenetic trees and machine learning classification techniques to reveal an underlying population structure.

Results

In this paper, we simulate phylogenetic trees from both structured and non-structured host populations. We compute eight statistics for the simulated trees, which are: the number of cherries; Sackin, Colless and total cophenetic indices; ladder length; maximum depth; maximum width, and width-to-depth ratio. Based on the estimated tree statistics, we classify the simulated trees as from either a non-structured or a structured population using the decision tree (DT), K-nearest neighbor (KNN) and support vector machine (SVM). We incorporate the basic reproductive number (\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$R_0$$\end{document}R0) in our tree simulation procedure. Sensitivity analysis is done to investigate whether the classifiers are robust to different choice of model parameters and to size of trees. Cross-validated results for area under the curve (AUC) for receiver operating characteristic (ROC) curves yield mean values of over 0.9 for most of the classification models.

Conclusions

Our classification procedure distinguishes well between trees from structured and non-structured populations using the classifiers, the two-sample Kolmogorov-Smirnov, Cucconi and Podgor-Gastwirth tests and the box plots. SVM models were more robust to changes in model parameters and tree size compared to KNN and DT classifiers. Our classification procedure was applied to real -world data and the structured population was revealed with high accuracy of \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$92.3\%$$\end{document}92.3% using SVM-polynomial classifier.

Model design for non-parametric phylodynamic inference and applications to pathogen surveillance

Inference of effective population size from genomic data can provide unique information about demographic history, and when applied to pathogen genetic data can also provide insights into epidemiological dynamics. The combination of non-parametric models for population dynamics with molecular clock models which relate genetic data to time has enabled phylodynamic inference based on large sets of time-stamped genetic sequence data. The methodology for non-parametric inference of effective population size is well-developed in the Bayesian setting, but here we develop a frequentist approach based on non-parametric latent process models of population size dynamics. We appeal to statistical principles based on out-of-sample prediction accuracy in order to optimize parameters that control shape and smoothness of the population size over time. We demonstrate the flexibility and speed of this approach in a series of simulation experiments, and apply the methodology to reconstruct the previously described waves in the seventh pandemic of cholera. We also estimate the impact of non-pharmaceutical interventions for COVID-19 in England using thousands of SARS-CoV-2 sequences. By incorporating a measure of the strength of these interventions over time within the phylodynamic model, we estimate the impact of the first national lockdown in the UK on the epidemic reproduction number.

Factors influencing estimates of HIV-1 infection timing using BEAST

While large datasets of HIV-1 sequences are increasingly being generated, many studies rely on a single gene or fragment of the genome and few comparative studies across genes have been done. We performed genome-based and gene-specific Bayesian phylogenetic analyses to investigate how certain factors impact estimates of the infection dates in an acute HIV-1 infection cohort, RV217. In this cohort, HIV-1 diagnosis corresponded to the first RNA positive test and occurred a median of four days after the last negative test, allowing us to compare timing estimates using BEAST to a narrow window of infection. We analyzed HIV-1 sequences sampled one week, one month and six months after HIV-1 diagnosis in 39 individuals. We found that shared diversity and temporal signal was limited in acute infection, and insufficient to allow timing inferences in the shortest HIV-1 genes, thus dated phylogenies were primarily analyzed for env, gag, pol and near full-length genomes. There was no one best-fitting model across participants and genes, though relaxed molecular clocks (73% of best-fitting models) and the Bayesian skyline (49%) tended to be favored. For infections with single founders, the infection date was estimated to be around one week pre-diagnosis for env (IQR: 3–9 days) and gag (IQR: 5–9 days), whilst the genome placed it at a median of 10 days (IQR: 4–19). Multiply-founded infections proved problematic to date. Our ability to compare timing inferences to precise estimates of HIV-1 infection (within a week) highlights that molecular dating methods can be applied to within-host datasets from early infection. Nonetheless, our results also suggest caution when using uniform clock and population models or short genes with limited information content.

Molecular dating using phylogenetics allows us to estimate the date of an infection from time-stamped within-host sequences alone. There are large datasets of HIV-1 sequences, but genome and gene analyses are not often performed in parallel and rarely with the possibility to compare results against a known narrow window of infection. We showed that all but the longest genes are near-clonal in acute infection, with little information for dating purposes. For infections with single founders, we estimated the eclipse phase—the time between HIV-1 exposure and the first positive diagnostic test—to last between one and two weeks using env, gag, pol and near full-length genomes. This approach could be used to narrow the date of suspected infection in ongoing clinical trials for the prevention of HIV-1 infection.

Agricultural and geographic factors shaped the North American 2015 highly pathogenic avian influenza H5N2 outbreak

The 2014-2015 highly pathogenic avian influenza (HPAI) H5NX outbreak represents the largest and most expensive HPAI outbreak in the United States to date. Despite extensive traditional and molecular epidemiological studies, factors associated with the spread of HPAI among midwestern poultry premises remain unclear. To better understand the dynamics of this outbreak, 182 full genome HPAI H5N2 sequences isolated from commercial layer chicken and turkey production premises were analyzed using evolutionary models able to accommodate epidemiological and geographic information. Epidemiological compartmental models embedded in a phylogenetic framework provided evidence that poultry type acted as a barrier to the transmission of virus among midwestern poultry farms. Furthermore, after initial introduction, the propagation of HPAI cases was self-sustainable within the commercial poultry industries. Discrete trait diffusion models indicated that within state viral transitions occurred more frequently than inter-state transitions. Distance and sample size were very strongly supported as associated with viral transition between county groups (Bayes Factor > 30.0). Together these findings indicate that the different types of midwestern poultry industries were not a single homogenous population, but rather, the outbreak was shaped by poultry industries and geographic factors.

Bayesian phylodynamic inference with complex models

Population genetic modeling can enhance Bayesian phylogenetic inference by providing a realistic prior on the distribution of branch lengths and times of common ancestry. The parameters of a population genetic model may also have intrinsic importance, and simultaneous estimation of a phylogeny and model parameters has enabled phylodynamic inference of population growth rates, reproduction numbers, and effective population size through time. Phylodynamic inference based on pathogen genetic sequence data has emerged as useful supplement to epidemic surveillance, however commonly-used mechanistic models that are typically fitted to non-genetic surveillance data are rarely fitted to pathogen genetic data due to a dearth of software tools, and the theory required to conduct such inference has been developed only recently. We present a framework for coalescent-based phylogenetic and phylodynamic inference which enables highly-flexible modeling of demographic and epidemiological processes. This approach builds upon previous structured coalescent approaches and includes enhancements for computational speed, accuracy, and stability. A flexible markup language is described for translating parametric demographic or epidemiological models into a structured coalescent model enabling simultaneous estimation of demographic or epidemiological parameters and time-scaled phylogenies. We demonstrate the utility of these approaches by fitting compartmental epidemiological models to Ebola virus and Influenza A virus sequence data, demonstrating how important features of these epidemics, such as the reproduction number and epidemic curves, can be gleaned from genetic data. These approaches are provided as an open-source package PhyDyn for the BEAST2 phylogenetics platform.

Modeling the Growth and Decline of Pathogen Effective Population Size Provides Insight into Epidemic Dynamics and Drivers of Antimicrobial Resistance

Nonparametric population genetic modeling provides a simple and flexible approach for studying demographic history and epidemic dynamics using pathogen sequence data. Existing Bayesian approaches are premised on stochastic processes with stationary increments which may provide an unrealistic prior for epidemic histories which feature extended period of exponential growth or decline. We show that nonparametric models defined in terms of the growth rate of the effective population size can provide a more realistic prior for epidemic history. We propose a nonparametric autoregressive model on the growth rate as a prior for effective population size, which corresponds to the dynamics expected under many epidemic situations. We demonstrate the use of this model within a Bayesian phylodynamic inference framework. Our method correctly reconstructs trends of epidemic growth and decline from pathogen genealogies even when genealogical data are sparse and conventional skyline estimators erroneously predict stable population size. We also propose a regression approach for relating growth rates of pathogen effective population size and time-varying variables that may impact the replicative fitness of a pathogen. The model is applied to real data from rabies virus and Staphylococcus aureus epidemics. We find a close correspondence between the estimated growth rates of a lineage of methicillin-resistant S. aureus and population-level prescription rates of \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{upgreek} \usepackage{mathrsfs} \setlength{\oddsidemargin}{-69pt} \begin{document} }{}$\beta$\end{document}-lactam antibiotics. The new models are implemented in an open source R package called skygrowth which is available at https://github.com/mrc-ide/skygrowth.

Recent advances in computational phylodynamics

Time-stamped, trait-annotated phylogenetic trees built from virus genome data are increasingly used for outbreak investigation and monitoring ongoing epidemics. This routinely involves reconstructing the spatial and demographic processes from large data sets to help unveil the patterns and drivers of virus spread. Such phylodynamic inferences can however become quite time-consuming as the dimensions of the data increase, which has led to a myriad of approaches that aim to tackle this complexity. To elucidate the current state of the art in the field of phylodynamics, we discuss recent developments in Bayesian inference and accompanying software, highlight methods for improving computational efficiency and relevant visualisation tools. As an alternative to fully Bayesian approaches, we touch upon conditional software pipelines that compromise between statistical coherence and turn-around-time, and we highlight the available software packages. Finally, we outline future directions that may facilitate the large-scale tracking of epidemics in near real time.

Phylodynamic Inference across Epidemic Scales

Within-host genetic diversity and large transmission bottlenecks confound phylodynamic inference of epidemiological dynamics. Conventional phylodynamic approaches assume that nodes in a time-scaled pathogen phylogeny correspond closely to the time of transmission between hosts that are ancestral to the sample. However, when hosts harbor diverse pathogen populations, node times can substantially pre-date infection times. Imperfect bottlenecks can cause lineages sampled in different individuals to coalesce in unexpected patterns. To address realistic violations of standard phylodynamic assumptions we developed a new inference approach based on a multi-scale coalescent model, accounting for nonlinear epidemiological dynamics, heterogeneous sampling through time, non-negligible genetic diversity of pathogens within hosts, and imperfect transmission bottlenecks. We apply this method to HIV-1 and Ebola virus (EBOV) outbreak sequence data, illustrating how and when conventional phylodynamic inference may give misleading results. Within-host diversity of HIV-1 causes substantial upwards bias in the number of infected hosts using conventional coalescent models, but estimates using the multi-scale model have greater consistency with reported number of diagnoses through time. In contrast, we find that within-host diversity of EBOV has little influence on estimated numbers of infected hosts or reproduction numbers, and estimates are highly consistent with the reported number of diagnoses through time. The multi-scale coalescent also enables estimation of within-host effective population size using single sequences from a random sample of patients. We find within-host population genetic diversity of HIV-1 p17 to be 2Nμ=0.012 (95% CI 0.0066-0.023), which is lower than estimates based on HIV envelope serial sequencing of individual patients.

Biased phylodynamic inferences from analysing clusters of viral sequences

Phylogenetic methods are being increasingly used to help understand the transmission dynamics of measurably evolving viruses, including HIV. Clusters of highly similar sequences are often observed, which appear to follow a ‘power law’ behaviour, with a small number of very large clusters. These clusters may help to identify subpopulations in an epidemic, and inform where intervention strategies should be implemented. However, clustering of samples does not necessarily imply the presence of a subpopulation with high transmission rates, as groups of closely related viruses can also occur due to non-epidemiological effects such as over-sampling. It is important to ensure that observed phylogenetic clustering reflects true heterogeneity in the transmitting population, and is not being driven by non-epidemiological effects. We qualify the effect of using a falsely identified ‘transmission cluster’ of sequences to estimate phylodynamic parameters including the effective population size and exponential growth rate under several demographic scenarios. Our simulation studies show that taking the maximum size cluster to re-estimate parameters from trees simulated under a randomly mixing, constant population size coalescent process systematically underestimates the overall effective population size. In addition, the transmission cluster wrongly resembles an exponential or logistic growth model 99% of the time. We also illustrate the consequences of false clusters in exponentially growing coalescent and birth-death trees, where again, the growth rate is skewed upwards. This has clear implications for identifying clusters in large viral databases, where a false cluster could result in wasted intervention resources.

Phylogenetic Tools for Generalized HIV-1 Epidemics: Findings from the PANGEA-HIV Methods Comparison

Viral phylogenetic methods contribute to understanding how HIV spreads in populations, and thereby help guide the design of prevention interventions. So far, most analyses have been applied to well-sampled concentrated HIV-1 epidemics in wealthy countries. To direct the use of phylogenetic tools to where the impact of HIV-1 is greatest, the Phylogenetics And Networks for Generalized HIV Epidemics in Africa (PANGEA-HIV) consortium generates full-genome viral sequences from across sub-Saharan Africa. Analyzing these data presents new challenges, since epidemics are principally driven by heterosexual transmission and a smaller fraction of cases is sampled. Here, we show that viral phylogenetic tools can be adapted and used to estimate epidemiological quantities of central importance to HIV-1 prevention in sub-Saharan Africa. We used a community-wide methods comparison exercise on simulated data, where participants were blinded to the true dynamics they were inferring. Two distinct simulations captured generalized HIV-1 epidemics, before and after a large community-level intervention that reduced infection levels. Five research groups participated. Structured coalescent modeling approaches were most successful: phylogenetic estimates of HIV-1 incidence, incidence reductions, and the proportion of transmissions from individuals in their first 3 months of infection correlated with the true values (Pearson correlation > 90%), with small bias. However, on some simulations, true values were markedly outside reported confidence or credibility intervals. The blinded comparison revealed current limits and strengths in using HIV phylogenetics in challenging settings, provided benchmarks for future methods' development, and supports using the latest generation of phylogenetic tools to advance HIV surveillance and prevention.

Genomic Analysis and Comparison of Two Gonorrhea Outbreaks

Gonorrhea is a sexually transmitted disease causing growing concern, with a substantial increase in reported incidence over the past few years in the United Kingdom and rising levels of resistance to a wide range of antibiotics. Understanding its epidemiology is therefore of major biomedical importance, not only on a population scale but also at the level of direct transmission. However, the molecular typing techniques traditionally used for gonorrhea infections do not provide sufficient resolution to investigate such fine-scale patterns. Here we sequenced the genomes of 237 isolates from two local collections of isolates from Sheffield and London, each of which was resolved into a single type using traditional methods. The two data sets were selected to have different epidemiological properties: the Sheffield data were collected over 6 years from a predominantly heterosexual population, whereas the London data were gathered within half a year and strongly associated with men who have sex with men. Based on contact tracing information between individuals in Sheffield, we found that transmission is associated with a median time to most recent common ancestor of 3.4 months, with an upper bound of 8 months, which we used as a criterion to identify likely transmission links in both data sets. In London, we found that transmission happened predominantly between individuals of similar age, sexual orientation, and location and also with the same HIV serostatus, which may reflect serosorting and associated risk behaviors. Comparison of the two data sets suggests that the London epidemic involved about ten times more cases than the Sheffield outbreak.

The recent increases in gonorrhea incidence and antibiotic resistance are cause for public health concern. Successful intervention requires a better understanding of transmission patterns, which is not uncovered by traditional molecular epidemiology techniques. Here we studied two outbreaks that took place in Sheffield and London, United Kingdom. We show that whole-genome sequencing provides the resolution to investigate direct gonorrhea transmission between infected individuals. Combining genome sequencing with rich epidemiological information about infected individuals reveals the importance of several transmission routes and risk factors, which can be used to design better control measures.

Phylogenetic Exploration of Nosocomial Transmission Chains of 2009 Influenza A/H1N1 among Children Admitted at Red Cross War Memorial Children's Hospital, Cape Town, South Africa in 2011

Traditional modes of investigating influenza nosocomial transmission have entailed a combination of confirmatory molecular diagnostic testing and epidemiological investigation. Common hospital-acquired infections like influenza require a discerning ability to distinguish between viral isolates to accurately identify patient transmission chains. We assessed whether influenza hemagglutinin sequence phylogenies can be used to enrich epidemiological data when investigating the extent of nosocomial transmission over a four-month period within a paediatric Hospital in Cape Town South Africa. Possible transmission chains/channels were initially determined through basic patient admission data combined with Maximum likelihood and time-scaled Bayesian phylogenetic analyses. These analyses suggested that most instances of potential hospital-acquired infections resulted from multiple introductions of Influenza A into the hospital, which included instances where virus hemagglutinin sequences were identical between different patients. Furthermore, a general inability to establish epidemiological transmission linkage of patients/viral isolates implied that identified isolates could have originated from asymptomatic hospital patients, visitors or hospital staff. In contrast, a traditional epidemiological investigation that used no viral phylogenetic analyses, based on patient co-admission into specific wards during a particular time-frame, suggested that multiple hospital acquired infection instances may have stemmed from a limited number of identifiable index viral isolates/patients. This traditional epidemiological analysis by itself could incorrectly suggest linkage between unrelated cases, underestimate the number of unique infections and may overlook the possible diffuse nature of hospital transmission, which was suggested by sequencing data to be caused by multiple unique introductions of influenza A isolates into individual hospital wards. We have demonstrated a functional role for viral sequence data in nosocomial transmission investigation through its ability to enrich traditional, non-molecular observational epidemiological investigation by teasing out possible transmission pathways and working toward more accurately enumerating the number of possible transmission events.

Measuring Asymmetry in Time-Stamped Phylogenies

Previous work has shown that asymmetry in viral phylogenies may be indicative of heterogeneity in transmission, for example due to acute HIV infection or the presence of ‘core groups’ with higher contact rates. Hence, evidence of asymmetry may provide clues to underlying population structure, even when direct information on, for example, stage of infection or contact rates, are missing. However, current tests of phylogenetic asymmetry (a) suffer from false positives when the tips of the phylogeny are sampled at different times and (b) only test for global asymmetry, and hence suffer from false negatives when asymmetry is localised to part of a phylogeny. We present a simple permutation-based approach for testing for asymmetry in a phylogeny, where we compare the observed phylogeny with random phylogenies with the same sampling and coalescence times, to reduce the false positive rate. We also demonstrate how profiles of measures of asymmetry calculated over a range of evolutionary times in the phylogeny can be used to identify local asymmetry. In combination with different metrics of asymmetry, this combined approach offers detailed insights of how phylogenies reconstructed from real viral datasets may deviate from the simplistic assumptions of commonly used coalescent and birth-death process models.

Phylogenetic trees of viruses sampled from different individuals provide clues to the dynamics of transmission. The extent to which the tree is asymmetric may be influenced by biological factors such as differences in infectiousness or contact rates between individuals, but also by nuisance factors such as the pattern of sampling. We have devised a simple statistical test for asymmetry, which controls for sampling patterns and potentially complex temporal dynamics by conditioning on the sampling and coalescence times in a phylogeny, and can also detect whether specific clades in the phylogeny drive patterns of asymmetry. We apply our approach to data on HIV, influenza A virus H5N1, and ebola virus.

Four key challenges in infectious disease modelling using data from multiple sources

Public health-related decision-making on policies aimed at controlling epidemics is increasingly evidence-based, exploiting multiple sources of data. Policy makers rely on complex models that are required to be robust, realistically approximating epidemics and consistent with all relevant data. Meeting these requirements in a statistically rigorous and defendable manner poses a number of challenging problems. How to weight evidence from different datasets and handle dependence between them, efficiently estimate and critically assess complex models are key challenges that we expound in this paper, using examples from influenza modelling.

Eight challenges in phylodynamic inference

The field of phylodynamics, which attempts to enhance our understanding of infectious disease dynamics using pathogen phylogenies, has made great strides in the past decade. Basic epidemiological and evolutionary models are now well characterized with inferential frameworks in place. However, significant challenges remain in extending phylodynamic inference to more complex systems. These challenges include accounting for evolutionary complexities such as changing mutation rates, selection, reassortment, and recombination, as well as epidemiological complexities such as stochastic population dynamics, host population structure, and different patterns at the within-host and between-host scales. An additional challenge exists in making efficient inferences from an ever increasing corpus of sequence data.

Inferring epidemiological dynamics with Bayesian coalescent inference: the merits of deterministic and stochastic models

Estimation of epidemiological and population parameters from molecular sequence data has become central to the understanding of infectious disease dynamics. Various models have been proposed to infer details of the dynamics that describe epidemic progression. These include inference approaches derived from Kingman’s coalescent theory. Here, we use recently described coalescent theory for epidemic dynamics to develop stochastic and deterministic coalescent susceptible–infected–removed (SIR) tree priors. We implement these in a Bayesian phylogenetic inference framework to permit joint estimation of SIR epidemic parameters and the sample genealogy. We assess the performance of the two coalescent models and also juxtapose results obtained with a recently published birth–death-sampling model for epidemic inference. Comparisons are made by analyzing sets of genealogies simulated under precisely known epidemiological parameters. Additionally, we analyze influenza A (H1N1) sequence data sampled in the Canterbury region of New Zealand and HIV-1 sequence data obtained from known United Kingdom infection clusters. We show that both coalescent SIR models are effective at estimating epidemiological parameters from data with large fundamental reproductive number R0 and large population size S0. Furthermore, we find that the stochastic variant generally outperforms its deterministic counterpart in terms of error, bias, and highest posterior density coverage, particularly for smaller R0 and S0. However, each of these inference models is shown to have undesirable properties in certain circumstances, especially for epidemic outbreaks with R0 close to one or with small effective susceptible populations.

Molecular epidemiology, phylogeny and evolution of the filarial nematode Wuchereria bancrofti

Wuchereria bancrofti (Wb) is the most widely distributed of the three nematodes known to cause lymphatic filariasis (LF), the other two being Brugia malayi and Brugia timori. Current tools available to monitor LF are limited to diagnostic tests targeting DNA repeats, filarial antigens, and anti-filarial antibodies. While these tools are useful for detection and surveillance, elimination programs have yet to take full advantage of molecular typing for inferring infection history, strain fingerprinting, and evolution. To date, molecular typing approaches have included whole mitochondrial genomes, genotyping, targeted sequencing, and random amplified polymorphic DNA (RAPDs). These studies have revealed much about Wb biology. For example, in one study in Papua New Guinea researchers identified 5 major strains that were widespread and many minor strains some of which exhibit geographic stratification. Genome data, while rare, has been utilized to reconstruct evolutionary relationships among taxa of the Onchocercidae (the clade of filarial nematodes) and identify gene synteny. Their phylogeny reveals that speciation from the common ancestor of both B. malayi and Wb occurred around 5-6 millions years ago with shared ancestry to other filarial nematodes as recent as 15 million years ago. These discoveries hold promise for gene discovery and identifying drug targets in species that are more amenable to in vivo experiments. Continued technological developments in whole genome sequencing and data analysis will likely replace many other forms of molecular typing, multiplying the amount of data available on population structure, genetic diversity, and phylogenetics. Once widely available, the addition of population genetic data from genomic studies should hasten the elimination of LF parasites like Wb. Infectious disease control programs have benefited greatly from population genetics data and recently from population genomics data. However, while there is currently a surplus of data for diseases like malaria and HIV, there is a scarcity of this data for filarial nematodes. With the falling cost of genome sequencing, research on filarial nematodes could benefit from the addition of population genetics statistics and phylogenetics especially in dealing with elimination programs. A comprehensive review focusing on population genetics of filarial nematode does not yet exist. Here our goal is to provide a current overview of the molecular epidemiology of W. bancrofti (Wb) the primary causative agent of LF. We begin by reviewing studies utilizing molecular typing techniques with specific focus on genomic and population datasets. Next, we used whole mitochondrial genome data to construct a phylogeny and examine the evolutionary history of the Onchocercidae. Then, we provide a perspective to aid in understanding how population genetic techniques translate to modern epidemiology. Finally, we introduce the concept of genomic epidemiology and provide some examples that will aid in future studies of Wb.

Phylodynamic inference for structured epidemiological models

Coalescent theory is routinely used to estimate past population dynamics and demographic parameters from genealogies. While early work in coalescent theory only considered simple demographic models, advances in theory have allowed for increasingly complex demographic scenarios to be considered. The success of this approach has lead to coalescent-based inference methods being applied to populations with rapidly changing population dynamics, including pathogens like RNA viruses. However, fitting epidemiological models to genealogies via coalescent models remains a challenging task, because pathogen populations often exhibit complex, nonlinear dynamics and are structured by multiple factors. Moreover, it often becomes necessary to consider stochastic variation in population dynamics when fitting such complex models to real data. Using recently developed structured coalescent models that accommodate complex population dynamics and population structure, we develop a statistical framework for fitting stochastic epidemiological models to genealogies. By combining particle filtering methods with Bayesian Markov chain Monte Carlo methods, we are able to fit a wide class of stochastic, nonlinear epidemiological models with different forms of population structure to genealogies. We demonstrate our framework using two structured epidemiological models: a model with disease progression between multiple stages of infection and a two-population model reflecting spatial structure. We apply the multi-stage model to HIV genealogies and show that the proposed method can be used to estimate the stage-specific transmission rates and prevalence of HIV. Finally, using the two-population model we explore how much information about population structure is contained in genealogies and what sample sizes are necessary to reliably infer parameters like migration rates.

Mathematical models play an important role in our understanding of what processes drive the complex population dynamics of infectious pathogens. Yet developing statistical methods for fitting models to epidemiological data is difficult. Epidemiological data is often noisy, incomplete, aggregated across different scales and generally provides only a partial picture of the underlying disease dynamics. Using nontraditional sources of data, like molecular sequences of pathogens, can provide additional information about epidemiological dynamics. But current “phylodynamic” inference methods for fitting models to genealogies reconstructed from sequence data have a number of major limitations. We present a statistical framework that builds upon earlier work to address two of these limitations: population structure and stochasticity. By incorporating population structure, our framework can be applied in cases where the host population is divided into different subpopulations, such as by spatial isolation. Our framework also takes into consideration stochastic noise and can therefore capture the inherent variability of epidemiological dynamics. These advances allow for a much wider class of epidemiological models to be fit to genealogies in order to estimate key epidemiological parameters and to reconstruct past disease dynamics.

Supersize me: how whole-genome sequencing and big data are transforming epidemiology

•

Whole-genome sequencing is used for forensic epidemiology.

•

Big data can transform forensic epidemiology.

•

Clustering, biases, wildlife reservoirs, and emerging infections can all be addressed.

•

Phylodynamics approaches to integrate epidemiological and evolutionary data have been highly successful but still face scientific challenges.

In epidemiology, the identification of ‘who infected whom’ allows us to quantify key characteristics such as incubation periods, heterogeneity in transmission rates, duration of infectiousness, and the existence of high-risk groups. Although invaluable, the existence of many plausible infection pathways makes this difficult, and epidemiological contact tracing either uncertain, logistically prohibitive, or both. The recent advent of next-generation sequencing technology allows the identification of traceable differences in the pathogen genome that are transforming our ability to understand high-resolution disease transmission, sometimes even down to the host-to-host scale. We review recent examples of the use of pathogen whole-genome sequencing for the purpose of forensic tracing of transmission pathways, focusing on the particular problems where evolutionary dynamics must be supplemented by epidemiological information on the most likely timing of events as well as possible transmission pathways. We also discuss potential pitfalls in the over-interpretation of these data, and highlight the manner in which a confluence of this technology with sophisticated mathematical and statistical approaches has the potential to produce a paradigm shift in our understanding of infectious disease transmission and control.

Phylogenetic analysis accounting for age-dependent death and sampling with applications to epidemics

The reconstruction of phylogenetic trees based on viral genetic sequence data sequentially sampled from an epidemic provides estimates of the past transmission dynamics, by fitting epidemiological models to these trees. To our knowledge, none of the epidemiological models currently used in phylogenetics can account for recovery rates and sampling rates dependent on the time elapsed since transmission, i.e. age of infection. Here we introduce an epidemiological model where infectives leave the epidemic, by either recovery or sampling, after some random time which may follow an arbitrary distribution. We derive an expression for the likelihood of the phylogenetic tree of sampled infectives under our general epidemiological model. The analytic concept developed in this paper will facilitate inference of past epidemiological dynamics and provide an analytical framework for performing very efficient simulations of phylogenetic trees under our model. The main idea of our analytic study is that the non-Markovian epidemiological model giving rise to phylogenetic trees growing vertically as time goes by can be represented by a Markovian "coalescent point process" growing horizontally by the sequential addition of pairs of coalescence and sampling times. As examples, we discuss two special cases of our general model, described in terms of influenza and HIV epidemics. Though phrased in epidemiological terms, our framework can also be used for instance to fit macroevolutionary models to phylogenies of extant and extinct species, accounting for general species lifetime distributions.

Simultaneous reconstruction of evolutionary history and epidemiological dynamics from viral sequences with the birth-death SIR model

The evolution of RNA viruses, such as human immunodeficiency virus (HIV), hepatitis C virus and influenza virus, occurs so rapidly that the viruses' genomes contain information on past ecological dynamics. Hence, we develop a phylodynamic method that enables the joint estimation of epidemiological parameters and phylogenetic history. Based on a compartmental susceptible–infected–removed (SIR) model, this method provides separate information on incidence and prevalence of infections. Detailed information on the interaction of host population dynamics and evolutionary history can inform decisions on how to contain or entirely avoid disease outbreaks. We apply our birth–death SIR method to two viral datasets. First, five HIV type 1 clusters sampled in the UK between 1999 and 2003 are analysed. The estimated basic reproduction ratios range from 1.9 to 3.2 among the clusters. All clusters show a decline in the growth rate of the local epidemic in the middle or end of the 1990s. The analysis of a hepatitis C virus genotype 2c dataset shows that the local epidemic in the Córdoban city Cruz del Eje originated around 1906 (median), coinciding with an immigration wave from Europe to central Argentina that dates from 1880 to 1920. The estimated time of epidemic peak is around 1970.

Reconciling phylodynamics with epidemiology: the case of dengue virus in southern Vietnam

Coalescent methods are widely used to infer the demographic history of populations from gene genealogies. These approaches—often referred to as phylodynamic methods—have proven especially useful for reconstructing the dynamics of rapidly evolving viral pathogens. Yet, population dynamics inferred from viral genealogies often differ widely from those observed from other sources of epidemiological data, such as hospitalization records. We demonstrate how a modeling framework that allows for the direct fitting of mechanistic epidemiological models to genealogies can be used to test different hypotheses about what ecological factors cause phylodynamic inferences to differ from observed dynamics. We use this framework to test different hypotheses about why dengue serotype 1 (DENV-1) population dynamics in southern Vietnam inferred using existing phylodynamic methods differ from hospitalization data. Specifically, we consider how factors such as seasonality, vector dynamics, and spatial structure can affect inferences drawn from genealogies. The coalescent models we derive to take into account vector dynamics and spatial structure reveal that these ecological complexities can substantially affect coalescent rates among lineages. We show that incorporating these additional ecological complexities into coalescent models can also greatly improve estimates of historical population dynamics and lead to new insights into the factors shaping viral genealogies.

Real-time characterization of the molecular epidemiology of an influenza pandemic

Early characterization of the epidemiology and evolution of a pandemic is essential for determining the most appropriate interventions. During the 2009 H1N1 influenza A pandemic, public databases facilitated widespread sharing of genetic sequence data from the outset. We use Bayesian phylogenetics to simulate real-time estimates of the evolutionary rate, date of emergence and intrinsic growth rate (r₀) of the pandemic from whole-genome sequences. We investigate the effects of temporal range of sampling and dataset size on the precision and accuracy of parameter estimation. Parameters can be accurately estimated as early as two months after the first reported case, from 100 genomes and the choice of growth model is important for accurate estimation of r₀. This demonstrates the utility of simple coalescent models to rapidly inform intervention strategies during a pandemic.