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

Exploring the Accuracy and Limits of Algorithms for Localizing Recombination Breakpoints

Phylogenetic methods are widely used to reconstruct the evolutionary relationships among species and individuals. However, recombination can obscure ancestral relationships as individuals may inherit different regions of their genome from different ancestors. It is, therefore, often necessary to detect recombination events, locate recombination breakpoints, and select recombination-free alignments prior to reconstructing phylogenetic trees. While many earlier studies have examined the power of different methods to detect recombination, very few have examined the ability of these methods to accurately locate recombination breakpoints. In this study, we simulated genome sequences based on ancestral recombination graphs and explored the accuracy of three popular recombination detection methods: MaxChi, 3SEQ, and Genetic Algorithm Recombination Detection. The accuracy of inferred breakpoint locations was evaluated along with the key factors contributing to variation in accuracy across datasets. While many different genomic features contribute to the variation in performance across methods, the number of informative sites consistent with the pattern of inheritance between parent and recombinant child sequences always has the greatest contribution to accuracy. While partitioning sequence alignments based on identified recombination breakpoints can greatly decrease phylogenetic error, the quality of phylogenetic reconstructions depends very little on how breakpoints are chosen to partition the alignment. Our work sheds light on how different features of recombinant genomes affect the performance of recombination detection methods and suggests best practices for reconstructing phylogenies based on recombination-free alignments.

Prevention of Viral Hepatitis and HIV Infection among People Who Inject Drugs: A Systematic Review and Meta-Analysis

This study aimed to explore the current evidence on preventing blood-borne virus infections among people who inject drugs (PWID). We conducted a comprehensive search across three databases (PubMed, Embase, Cochrane Library) for relevant articles published in English between 2014 and 2023. We followed the Preferred Reporting Items for Systematic Reviews and Meta Analysis (PRISMA) guidelines, assessed the quality of the paper using the revised Cochrane Risk of Bias Tool (ROB 2), and conducted a meta-analysis using RevMan 5.3. Completing the harm reduction program (HRP) participation and receiving all three vaccine doses resulted in a 28% reduction in the risk of HBV infection (OR: 0.72, 95% CI: 0.37-1.42). Various interventions increased the willingness of PWIDs to undergo HCV treatment (OR: 5.91, 95% CI: 2.46-14.24) and promoted treatment adherence (OR: 15.04, 95% CI: 2.80-80.61). Taking PrEP, participating in HRP, and modifying risky behaviors were associated with a 33% reduction in the risk of HIV infection (OR: 0.67, 95% CI: 0.61-0.74). Conducting referrals, providing counseling, and implementing antiretroviral therapy resulted in a 44% reduction in the risk of viral transmission (OR: 0.56, 95% CI: 0.47-0.66). Co-infection may potentially compromise effectiveness, so it is important to consider drug resistance.

Epidemiological inference for emerging viruses using segregating sites

Epidemiological models are commonly fit to case and pathogen sequence data to estimate parameters and to infer unobserved disease dynamics. Here, we present an inference approach based on sequence data that is well suited for model fitting early on during the expansion of a viral lineage. Our approach relies on a trajectory of segregating sites to infer epidemiological parameters within a Sequential Monte Carlo framework. Using simulated data, we first show that our approach accurately recovers key epidemiological quantities under a single-introduction scenario. We then apply our approach to SARS-CoV-2 sequence data from France, estimating a basic reproduction number of approximately 2.3-2.7 under an epidemiological model that allows for multiple introductions. Our approach presented here indicates that inference approaches that rely on simple population genetic summary statistics can be informative of epidemiological parameters and can be used for reconstructing infectious disease dynamics during the early expansion of a viral lineage.

Community Insights in Phylogenetic HIV Research: The CIPHR Project Protocol

Inferring HIV transmission networks from HIV sequences is gaining popularity in the field of HIV molecular epidemiology. However, HIV sequences are often analyzed at distance from those affected by HIV epidemics, namely without the involvement of communities most affected by HIV. These remote analyses often mean that knowledge is generated in absence of lived experiences and socio-economic realities that could inform the ethical application of network-derived information in 'real world' programmes. Procedures to engage communities are noticeably absent from the HIV molecular epidemiology literature. Here we present our team's protocol for engaging community activists living in Nairobi, Kenya in a knowledge exchange process - The CIPHR Project (Community Insights in Phylogenetic HIV Research). Drawing upon a community-based participatory approach, our team will (1) explore the possibilities and limitations of HIV molecular epidemiology for key population programmes, (2) pilot a community-based HIV molecular study, and (3) co-develop policy guidelines on conducting ethically safe HIV molecular epidemiology. Critical dialogue with activist communities will offer insight into the potential uses and abuses of using such information to sharpen HIV prevention programmes. The outcome of this process holds importance to the development of policy frameworks that will guide the next generation of the global response.

A scalable model for simulating multi-round antibody evolution and benchmarking of clonal tree reconstruction methods

Affinity maturation (AM) of B cells through somatic hypermutations (SHMs) enables the immune system to evolve to recognize diverse pathogens. The accumulation of SHMs leads to the formation of clonal lineages of antibody-secreting b cells that have evolved from a common naïve B cell. Advances in high-throughput sequencing have enabled deep scans of B cell receptor repertoires, paving the way for reconstructing clonal trees. However, it is not clear if clonal trees, which capture microevolutionary time scales, can be reconstructed using traditional phylogenetic reconstruction methods with adequate accuracy. In fact, several clonal tree reconstruction methods have been developed to fix supposed shortcomings of phylogenetic methods. Nevertheless, no consensus has been reached regarding the relative accuracy of these methods, partially because evaluation is challenging. Benchmarking the performance of existing methods and developing better methods would both benefit from realistic models of clonal lineage evolution specifically designed for emulating B cell evolution. In this paper, we propose a model for modeling B cell clonal lineage evolution and use this model to benchmark several existing clonal tree reconstruction methods. Our model, designed to be extensible, has several features: by evolving the clonal tree and sequences simultaneously, it allows modeling selective pressure due to changes in affinity binding; it enables scalable simulations of large numbers of cells; it enables several rounds of infection by an evolving pathogen; and, it models building of memory. In addition, we also suggest a set of metrics for comparing clonal trees and measuring their properties. Our results show that while maximum likelihood phylogenetic reconstruction methods can fail to capture key features of clonal tree expansion if applied naively, a simple post-processing of their results, where short branches are contracted, leads to inferences that are better than alternative methods.

PopART-IBM, a highly efficient stochastic individual-based simulation model of generalised HIV epidemics developed in the context of the HPTN 071 (PopART) trial

Mathematical models are powerful tools in HIV epidemiology, producing quantitative projections of key indicators such as HIV incidence and prevalence. In order to improve the accuracy of predictions, such models need to incorporate a number of behavioural and biological heterogeneities, especially those related to the sexual network within which HIV transmission occurs. An individual-based model, which explicitly models sexual partnerships, is thus often the most natural type of model to choose. In this paper we present PopART-IBM, a computationally efficient individual-based model capable of simulating 50 years of an HIV epidemic in a large, high-prevalence community in under a minute. We show how the model calibrates within a Bayesian inference framework to detailed age- and sex-stratified data from multiple sources on HIV prevalence, awareness of HIV status, ART status, and viral suppression for an HPTN 071 (PopART) study community in Zambia, and present future projections of HIV prevalence and incidence for this community in the absence of trial intervention.

In this paper we present PopART-IBM, an individual-based model used to simulate HIV transmission in communities in high prevalence settings. We show that PopART-IBM can simulate transmission over a span of decades in a large community in less than a minute. This computational efficiency allows us to calibrate the model within an inference framework, and we show an illustrative example of calibration using an adaptive population Monte Carlo Approximate Bayesian Computation algorithm for a community in Zambia that was part of the HPTN-071 (PopART) trial. We compare the detailed model output to real-world data collected during the trial from this community. Finally, we project how the HIV epidemic would have changed over time in this community if no intervention from the trial had occurred.

SEPIA: simulation-based evaluation of prioritization algorithms

Background

The ability to prioritize people living with HIV (PLWH) by risk of future transmissions could aid public health officials in optimizing epidemiological intervention. While methods exist to perform such prioritization based on molecular data, their effectiveness and accuracy are poorly understood, and it is unclear how one can directly compare the accuracy of different methods. We introduce SEPIA (Simulation-based Evaluation of PrIoritization Algorithms), a novel simulation-based framework for determining the effectiveness of prioritization algorithms. SEPIA expands upon prior related work by defining novel metrics of effectiveness with which to compare prioritization techniques, as well as by creating a simulation-based tool with which to perform such effectiveness comparisons. Under several metrics of effectiveness that we propose, we compare two existing prioritization approaches: one phylogenetic (ProACT) and one distance-based (growth of HIV-TRACE transmission clusters).

Results

Using all proposed metrics, ProACT consistently slightly outperformed the transmission cluster growth approach. However, both methods consistently performed just marginally better than random, suggesting that there is significant room for improvement in prioritization tools.

Conclusion

We hope that, by providing ways to quantify the effectiveness of prioritization methods in simulation, SEPIA will aid researchers in developing novel risk prioritization tools for PLWH.

Timing the SARS-CoV-2 index case in Hubei province

Understanding when severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) emerged is critical to evaluating our current approach to monitoring novel zoonotic pathogens and understanding the failure of early containment and mitigation efforts for COVID-19. We used a coalescent framework to combine retrospective molecular clock inference with forward epidemiological simulations to determine how long SARS-CoV-2 could have circulated before the time of the most recent common ancestor of all sequenced SARS-CoV-2 genomes. Our results define the period between mid-October and mid-November 2019 as the plausible interval when the first case of SARS-CoV-2 emerged in Hubei province, China. By characterizing the likely dynamics of the virus before it was discovered, we show that more than two-thirds of SARS-CoV-2-like zoonotic events would be self-limited, dying out without igniting a pandemic. Our findings highlight the shortcomings of zoonosis surveillance approaches for detecting highly contagious pathogens with moderate mortality rates.

HIV Care Prioritization Using Phylogenetic Branch Length

BACKGROUND

The structure of the HIV transmission networks can be dictated by just a few individuals. Public health intervention, such as ensuring people living with HIV adhere to antiretroviral therapy and remain virally suppressed, can help control the spread of the virus. However, such intervention requires using limited public health resource allocations. Determining which individuals are most at risk of transmitting HIV could allow public health officials to focus their limited resources on these individuals.

SETTING

Molecular epidemiology can help prioritize people living with HIV by patterns of transmission inferred from their sampled viral sequences. Such prioritization has been previously suggested and performed by monitoring cluster growth. In this article, we introduce Prioritization using AnCesTral edge lengths (ProACT), a phylogenetic approach for prioritizing individuals living with HIV.

METHODS

ProACT starts from a phylogeny inferred from sequence data and orders individuals according to their terminal branch length, breaking ties using ancestral branch lengths. We evaluated ProACT on a real data set of 926 HIV-1 subtype B pol data obtained in San Diego between 2005 and 2014 and a simulation data set modeling the same epidemic. Prioritization methods are compared by their ability to predict individuals who transmit most after the prioritization.

RESULTS

Across all simulation conditions and most real data sampling conditions, ProACT outperformed monitoring cluster growth for multiple metrics of prioritization efficacy.

CONCLUSION

The simple strategy used by ProACT improves the effectiveness of prioritization compared with state-of-the-art methods that rely on monitoring the growth of transmission clusters defined based on genetic distance.

Field validation of phylodynamic analytical methods for inference on epidemiological processes in wildlife

Amongst newly developed approaches to analyse molecular data, phylodynamic models are receiving much attention because of their potential to reveal changes to viral populations over short periods. This knowledge can be very important for understanding disease impacts. However, their accuracy needs to be fully understood, especially in relation to wildlife disease epidemiology, where sampling and prior knowledge may be limited. The release of the rabbit haemorrhagic disease virus (RHDV) as biological control in naïve rabbit populations in Australia in 1996 provides a unique data set with which to validate phylodynamic models. By comparing results obtained from RHDV sequence data with our current understanding of RHDV epidemiology in Australia, we evaluated the performances of these recently developed models. In line with our expectations, coalescent analyses detected a sharp increase in the virus population size in the first few months after release, followed by a more gradual increase. Phylodynamic analyses using a birth-death model generated effective reproductive number estimates (the average number of secondary infections per each infectious case, R_e ) larger than one for most of the epochs considered. However, the possible range of the initial R_e included estimates lower than one despite the known rapid spread of RHDV in Australia. Furthermore, the analyses that accounted for geographical structuring failed to converge. We argue that the difficulties that we encountered most likely stem from the fact that the samples available from 1996 to 2014 were too sparse with respect to both geographic and within outbreak coverage to adequately infer some of the model parameters. In general, while these phylodynamic analyses proved to be greatly informative in some regards, we caution that their interpretation may not be straightforward. We recommend further research to evaluate the robustness of these models to assumption violations and sensitivity to sampling regimes.

Effective HIV Molecular Surveillance Requires Identification of Incident Cases of Infection

BACKGROUND

Ending the HIV epidemic requires knowledge of key drivers of spread of HIV infection.

METHODS

Between 1996 and 2018, 1119 newly and previously diagnosed, therapy-naïve persons with HIV (PWH) from San Diego were followed. A genetic distance-based network was inferred using pol sequences, and genetic clusters grew over time through linkage of sequences from newly observed infections. Cox proportional hazards models were used to identify factors associated with the rate of growth. These results were used to predict the impact of a hypothetical intervention targeting PWH with incident infection. Comparison was made to the CDC EHE molecular surveillance strategy, which prioritizes clusters recently linked to all new HIV diagnoses and does not incorporate data on incident infections.

RESULTS

Overall, 219 genetic linkages to incident infections were identified over a median follow-up of 8.8 years. Incident cluster growth was strongly associated with proportion of PWH in the cluster who themselves had incident infection. (HR 44.09; 95% CI: 17.09, 113.78). The CDC EHE molecular surveillance strategy identified 11 linkages to incident infections a genetic distance threshold of 0.5%, and 24 linkages at 1.5%.

CONCLUSIONS

Over the past two decades, incident infections drove incident HIV cluster growth in San Diego. The current CDC EHE molecular detection and response strategy would not have identified most transmission events arising from those with incident infection in San Diego. Molecular surveillance that includes detection of incident cases will provide a more effective strategy for EHE.

Genomic Epidemiology Analysis of Infectious Disease Outbreaks Using TransPhylo

Comparing the pathogen genomes from several cases of an infectious disease has the potential to help us understand and control outbreaks. Many methods exist to reconstruct a phylogeny from such genomes, which represents how the genomes are related to one another. However, such a phylogeny is not directly informative about transmission events between individuals. TransPhylo is a software tool implemented as an R package designed to bridge the gap between pathogen phylogenies and transmission trees. TransPhylo is based on a combined model of transmission between hosts and pathogen evolution within each host. It can simulate both phylogenies and transmission trees jointly under this combined model. TransPhylo can also reconstruct a transmission tree based on a dated phylogeny, by exploring the space of transmission trees compatible with the phylogeny. A transmission tree can be represented as a coloring of a phylogeny where each color represents a different host of the pathogen, and TransPhylo provides convenient ways to plot these colorings and explore the results. This article presents the basic protocols that can be used to make the most of TransPhylo. © 2021 The Authors. Basic Protocol 1: First steps with TransPhylo Basic Protocol 2: Simulation of outbreak data Basic Protocol 3: Inference of transmission Basic Protocol 4: Exploring the results of inference.

The probability distribution of the ancestral population size conditioned on the reconstructed phylogenetic tree with occurrence data

We consider a homogeneous birth-death process with three different sampling schemes. First, individuals can be sampled through time and included in a reconstructed phylogenetic tree. Second, they can be sampled through time and only recorded as a point 'occurrence' along a timeline. Third, extant individuals can be sampled and included in the reconstructed phylogenetic tree with a fixed probability. We further consider that sampled individuals can be removed or not from the process, upon sampling, with fixed probability. We derive the probability distribution of the population size at any time in the past conditional on the joint observation of a reconstructed phylogenetic tree and a record of occurrences not included in the tree. We also provide an algorithm to simulate ancestral population size trajectories given the observation of a reconstructed phylogenetic tree and occurrences. This distribution can be readily used to draw inferences about the ancestral population size in the field of epidemiology and macroevolution. In epidemiology, these results will allow data from epidemiological case count studies to be used in conjunction with molecular sequencing data (yielding reconstructed phylogenetic trees) to coherently estimate prevalence through time. In macroevolution, it will foster the joint examination of the fossil record and extant taxa to reconstruct past biodiversity.

Empirical comparison of analytical approaches for identifying molecular HIV-1 clusters

Public health interventions guided by clustering of HIV-1 molecular sequences may be impacted by choices of analytical approaches. We identified commonly-used clustering analytical approaches, applied them to 1886 HIV-1 Rhode Island sequences from 2004-2018, and compared concordance in identifying molecular HIV-1 clusters within and between approaches. We used strict (topological support ≥ 0.95; distance 0.015 substitutions/site) and relaxed (topological support 0.80-0.95; distance 0.030-0.045 substitutions/site) thresholds to reflect different epidemiological scenarios. We found that clustering differed by method and threshold and depended more on distance than topological support thresholds. Clustering concordance analyses demonstrated some differences across analytical approaches, with RAxML having the highest (91%) mean summary percent concordance when strict thresholds were applied, and three (RAxML-, FastTree regular bootstrap- and IQ-Tree regular bootstrap-based) analytical approaches having the highest (86%) mean summary percent concordance when relaxed thresholds were applied. We conclude that different analytical approaches can yield diverse HIV-1 clustering outcomes and may need to be differentially used in diverse public health scenarios. Recognizing the variability and limitations of commonly-used methods in cluster identification is important for guiding clustering-triggered interventions to disrupt new transmissions and end the HIV epidemic.

What Should Health Departments Do with HIV Sequence Data?

Many countries and US states have mandatory statues that require reporting of HIV clinical data including genetic sequencing results to the public health departments. Because genetic sequencing is a part of routine care for HIV infected persons, health departments have extensive sequence collections spanning years and even decades of the HIV epidemic. How should these data be used (or not) in public health practice? This is a complex, multi-faceted question that weighs personal risks against public health benefit. The answer is neither straightforward nor universal. However, to make that judgement—of how genetic sequence data should be used in describing and combating the HIV epidemic—we need a clear image of what a phylogenetically enhanced HIV surveillance system can do and what benefit it might provide. In this paper, we present a positive case for how up-to-date analysis of HIV sequence databases managed by health departments can provide unique and actionable information of how HIV is spreading in local communities. We discuss this question broadly, with examples from the US, as it is globally relevant for all health authorities that collect HIV genetic data.

Number of HIV-1 founder variants is determined by the recency of the source partner infection

During sexual transmission, the high genetic diversity of HIV-1 within an individual is frequently reduced to one founder variant that initiates infection. Understanding the drivers of this bottleneck is crucial to developing effective infection control strategies. Little is known about the importance of the source partner during this bottleneck. To test the hypothesis that the source partner affects the number of HIV founder variants, we developed a phylodynamic model calibrated using genetic and epidemiological data on all existing transmission pairs for whom the direction of transmission and the infection stage of the source partner are known. Our results suggest that acquiring infection from someone in the acute (early) stage of infection increases the risk of multiple-founder variant transmission compared with acquiring infection from someone in the chronic (later) stage of infection. This study provides the first direct test of source partner characteristics to explain the low frequency of multiple-founder strain infections.

Validation of Variant Assembly Using HAPHPIPE with Next-Generation Sequence Data from Viruses

Next-generation sequencing (NGS) offers a powerful opportunity to identify low-abundance, intra-host viral sequence variants, yet the focus of many bioinformatic tools on consensus sequence construction has precluded a thorough analysis of intra-host diversity. To take full advantage of the resolution of NGS data, we developed HAplotype PHylodynamics PIPEline (HAPHPIPE), an open-source tool for the de novo and reference-based assembly of viral NGS data, with both consensus sequence assembly and a focus on the quantification of intra-host variation through haplotype reconstruction. We validate and compare the consensus sequence assembly methods of HAPHPIPE to those of two alternative software packages, HyDRA and Geneious, using simulated HIV and empirical HIV, HCV, and SARS-CoV-2 datasets. Our validation methods included read mapping, genetic distance, and genetic diversity metrics. In simulated NGS data, HAPHPIPE generated pol consensus sequences significantly closer to the true consensus sequence than those produced by HyDRA and Geneious and performed comparably to Geneious for HIV gp120 sequences. Furthermore, using empirical data from multiple viruses, we demonstrate that HAPHPIPE can analyze larger sequence datasets due to its greater computational speed. Therefore, we contend that HAPHPIPE provides a more user-friendly platform for users with and without bioinformatics experience to implement current best practices for viral NGS assembly than other currently available options.

FAVITES: simultaneous simulation of transmission networks, phylogenetic trees and sequences

MOTIVATION

The ability to simulate epidemics as a function of model parameters allows insights that are unobtainable from real datasets. Further, reconstructing transmission networks for fast-evolving viruses like Human Immunodeficiency Virus (HIV) may have the potential to greatly enhance epidemic intervention, but transmission network reconstruction methods have been inadequately studied, largely because it is difficult to obtain 'truth' sets on which to test them and properly measure their performance.

RESULTS

We introduce FrAmework for VIral Transmission and Evolution Simulation (FAVITES), a robust framework for simulating realistic datasets for epidemics that are caused by fast-evolving pathogens like HIV. FAVITES creates a generative model to produce contact networks, transmission networks, phylogenetic trees and sequence datasets, and to add error to the data. FAVITES is designed to be extensible by dividing the generative model into modules, each of which is expressed as a fixed API that can be implemented using various models. We use FAVITES to simulate HIV datasets and study the realism of the simulated datasets. We then use the simulated data to study the impact of the increased treatment efforts on epidemiological outcomes. We also study two transmission network reconstruction methods and their effectiveness in detecting fast-growing clusters.

AVAILABILITY AND IMPLEMENTATION

FAVITES is available at https://github.com/niemasd/FAVITES, and a Docker image can be found on DockerHub (https://hub.docker.com/r/niemasd/favites).

SUPPLEMENTARY INFORMATION

Supplementary data are available at Bioinformatics online.

Recombination Analysis of Near Full-Length HIV-1 Sequences and the Identification of a Potential New Circulating Recombinant Form from Rakai, Uganda

The Phylogenetics And Networks for Generalized HIV Epidemics in Africa (PANGEA-HIV) consortium has been vital in the generation and examination of near full-length HIV-1 sequences generated from Sub-Saharan Africa. In this study, we examined a subset (n = 275) of sequences from Rakai, Uganda, collected between August 2011 and January 2015. Sequences were initially screened with COMET for subtyping and then evaluated using bootscanning and phylogenetic inference. Among 275 sequences, 38.6% were subtype D, 19.3% were subtype A, 2.9% were subtype C, and 39.3% were recombinant. The recombinants were structurally diverse in the number of breakpoints observed, the location of recombinant segments, and represented subtypes, with AD recombinants accounting for the majority of all recombinants (29.8%). Within the AD subpopulation, we identified a potential new circulating recombinant form in five individuals where the polymerase gene was subtype D and most of env was subtype A (D-A junctures at HXB2 6760 and 8709). While the breakpoints were identical for the viruses from these individuals, the viral fragments did not cluster together. These results suggest selection for a viral strain where properties of the subtype A and subtype D portions of the virus confer a survival advantage. The continued study of recombinants will increase our breadth of knowledge for the genetic diversity and evolution of HIV-1, which can further contribute to our understanding toward a universal HIV-1 vaccine.

SimpactCyan 1.0: An Open-source Simulator for Individual-Based Models in HIV Epidemiology with R and Python Interfaces

SimpactCyan is an open-source simulator for individual-based models in HIV epidemiology. Its core algorithm is written in C++ for computational efficiency, while the R and Python interfaces aim to make the tool accessible to the fast-growing community of R and Python users. Transmission, treatment and prevention of HIV infections in dynamic sexual networks are simulated by discrete events. A generic “intervention” event allows model parameters to be changed over time, and can be used to model medical and behavioural HIV prevention programmes. First, we describe a more efficient variant of the modified Next Reaction Method that drives our continuous-time simulator. Next, we outline key built-in features and assumptions of individual-based models formulated in SimpactCyan, and provide code snippets for how to formulate, execute and analyse models in SimpactCyan through its R and Python interfaces. Lastly, we give two examples of applications in HIV epidemiology: the first demonstrates how the software can be used to estimate the impact of progressive changes to the eligibility criteria for HIV treatment on HIV incidence. The second example illustrates the use of SimpactCyan as a data-generating tool for assessing the performance of a phylodynamic inference framework.

Phylodynamic analysis of HIV-1 subtypes B, C and CRF 02_AG in Senegal

Surveillance of HIV epidemics in key populations and in developing countries is often challenging due to sparse, incomplete, or low-quality data. Analysis of HIV sequence data can provide an alternative source of information about epidemic history, population structure, and transmission patterns. To understand HIV-1 dynamics and transmission patterns in Senegal, we carried out model-based phylodynamic analyses using the structured-coalescent approach using HIV-1 sequence data from three different subgroups: reproductive aged males and females from the adult Senegalese population and men who have sex with other men (MSM). We fitted these phylodynamic analyses to time-scaled phylogenetic trees individually for subtypes C and CRF 02_AG, and for the combined data for subtypes B, C and CRF 02_AG. In general, the combined analysis showed a decreasing proportion of effective number of infections among all reproductive aged adults relative to MSM. However, we observed a nearly time-invariant distribution for subtype CRF 02_AG and an increasing trend for subtype C on the proportion of effective number of infections. The population attributable fraction also differed between analyses: subtype CRF 02_AG showed little contribution from MSM, while for subtype C and combined analyses this contribution was much higher. Despite observed differences, results suggested that the combination of high assortativity among MSM and the unmet HIV prevention and treatment needs represent a significant component of the HIV epidemic in Senegal.

A non-parametric analytic framework for within-host viral phylogenies and a test for HIV-1 founder multiplicity

Phylogenetics is a powerful tool for understanding the diversification dynamics of viral pathogens. Here we present an extension of the spectral density profile of the modified graph Laplacian, which facilitates the characterization of within-host molecular evolution of viruses and the direct comparison of diversification dynamics between hosts. This approach is non-parametric and therefore fast and model-free. We used simulations of within-host evolutionary scenarios to evaluate the efficiency of our approach and to demonstrate the significance of interpreting a viral phylogeny by its spectral density profile in terms of diversification dynamics. The key features that are captured by the profile are positive selection on the viral gene (or genome), temporal changes in substitution rates, mutational fitness, and time between sampling. Using sequences from individuals infected with HIV-1, we showed the utility of this approach for characterizing within-host diversification dynamics, for comparing dynamics between hosts, and for charting disease progression in infected individuals sampled over multiple years. We furthermore propose a heuristic test for assessing founder heterogeneity, which allows us to classify infections with single and multiple HIV-1 founder viruses. This non-parametric approach can be a valuable complement to existing parametric approaches.

Detection of HIV transmission hotspots in British Columbia, Canada: A novel framework for the prioritization and allocation of treatment and prevention resources

Background

Identifying populations at high risk of HIV transmission is critical for prioritizing treatment and prevention resources and achieving the UNAIDS 90-90-90 Targets.

Methods

HIV transmission rates can be estimated from phylogenetic trees as viral lineage-level diversification rates. To identify HIV-1 transmission foci in British Columbia, Canada, we inferred diversification rates from phylogenetic trees of 36 271 HIV-1 sequences from 9630 anonymized individuals. Diversification rates were combined with sociodemographic and clinical data, then aggregated by patients’ area of residence to predict the distribution of new HIV cases between 2008 and 2018. The predictive power of the model was compared with a phylogenetically uninformed model.

Findings

Aggregated diversification rate measures were predictive of new HIV cases in the subsequent year after adjusting for prevalent and incident cases in the previous year. For every one-unit increase in the mean of the top five diversification rates, the number of new HIV cases increased by on average 1·38-fold (95% CI, 1·28–1·49). In a blind prediction of 2018 cases, diversification rate improved the model's specificity by 12%, accuracy by 9%, top 20 agreement by 100%, and correlation of predicted and observed values by 162% relative to a model that incorporated epidemiological data alone.

Interpretation

By predicting the distribution of future HIV cases, a combined phylogenetic and epidemiological approach identifies hotspots where public health resources are needed most.

Funding

Canadian Institutes of Health Research, University of British Columbia, Public Health Agency of Canada, Genome Canada, Genome BC, Michael Smith Foundation for Health Research, and BC Centre for Excellence in HIV/AIDS.

Assessing biases in phylodynamic inferences in the presence of super-spreaders

Phylodynamic analyses using pathogen genetic data have become popular for making epidemiological inferences. However, many methods assume that the underlying host population follows homogenous mixing patterns. Nevertheless, in real disease outbreaks, a small number of individuals infect a disproportionately large number of others (super-spreaders). Our objective was to quantify the degree of bias in estimating the epidemic starting date in the presence of super-spreaders using different sample selection strategies. We simulated 100 epidemics of a hypothetical pathogen (fast evolving foot and mouth disease virus-like) over a real livestock movement network allowing the genetic mutations in pathogen sequence. Genetic sequences were sampled serially over the epidemic, which were then used to estimate the epidemic starting date using Extended Bayesian Coalescent Skyline plot (EBSP) and Birth–death skyline plot (BDSKY) models. Our results showed that the degree of bias varies over different epidemic situations, with substantial overestimations on the epidemic duration occurring in some occasions. While the accuracy and precision of BDSKY were deteriorated when a super-spreader generated a larger proportion of secondary cases, those of EBSP were deteriorated when epidemics were shorter. The accuracies of the inference were similar irrespective of whether the analysis used all sampled sequences or only a subset of them, although the former required substantially longer computational times. When phylodynamic analyses need to be performed under a time constraint to inform policy makers, we suggest multiple phylodynamics models to be used simultaneously for a subset of data to ascertain the robustness of inferences.

Molecular Epidemiology of HIV-1 Subtype B Reveals Heterogeneous Transmission Risk: Implications for Intervention and Control

Background

The impact of HIV pre-exposure prophylaxis (PrEP) depends on infections averted by protecting vulnerable individuals as well as infections averted by preventing transmission by those who would have been infected if not receiving PrEP. Analysis of HIV phylogenies reveals risk factors for transmission, which we examine as potential criteria for allocating PrEP.

Methods

We analyzed 6912 HIV-1 partial pol sequences from men who have sex with men (MSM) in the United Kingdom combined with global reference sequences and patient-level metadata. Population genetic models were developed that adjust for stage of infection, global migration of HIV lineages, and changing incidence of infection through time. Models were extended to simulate the effects of providing susceptible MSM with PrEP.

Results

We found that young age <25 years confers higher risk of HIV transmission (relative risk = 2.52 [95% confidence interval, 2.32–2.73]) and that young MSM are more likely to transmit to one another than expected by chance. Simulated interventions indicate that 4-fold more infections can be averted over 5 years by focusing PrEP on young MSM.

Conclusions

Concentrating PrEP doses on young individuals can avert more infections than random allocation.

A Fast Likelihood Method to Reconstruct and Visualize Ancestral Scenarios

The reconstruction of ancestral scenarios is widely used to study the evolution of characters along phylogenetic trees. One commonly uses the marginal posterior probabilities of the character states, or the joint reconstruction of the most likely scenario. However, marginal reconstructions provide users with state probabilities, which are difficult to interpret and visualize, whereas joint reconstructions select a unique state for every tree node and thus do not reflect the uncertainty of inferences. We propose a simple and fast approach, which is in between these two extremes. We use decision-theory concepts (namely, the Brier score) to associate each node in the tree to a set of likely states. A unique state is predicted in tree regions with low uncertainty, whereas several states are predicted in uncertain regions, typically around the tree root. To visualize the results, we cluster the neighboring nodes associated with the same states and use graph visualization tools. The method is implemented in the PastML program and web server. The results on simulated data demonstrate the accuracy and robustness of the approach. PastML was applied to the phylogeography of Dengue serotype 2 (DENV2), and the evolution of drug resistances in a large HIV data set. These analyses took a few minutes and provided convincing results. PastML retrieved the main transmission routes of human DENV2 and showed the uncertainty of the human-sylvatic DENV2 geographic origin. With HIV, the results show that resistance mutations mostly emerge independently under treatment pressure, but resistance clusters are found, corresponding to transmissions among untreated patients.

PANGEA-HIV 2: Phylogenetics And Networks for Generalised Epidemics in Africa

PURPOSE OF REVIEW

The HIV epidemic in sub-Saharan Africa is far from being under control and the ambitious UNAIDS targets are unlikely to be met by 2020 as declines in per-capita incidence being largely offset by demographic trends. There is an increasing number of proven and specific HIV prevention tools, but little consensus on how best to deploy them.

RECENT FINDINGS

Traditionally, phylogenetics has been used in HIV research to reconstruct the history of the epidemic and date zoonotic infections, whereas more recent publications focus on HIV diversity and drug resistance. However, it is also the most powerful method of source attribution available for the study of HIV transmission. The PANGEA (Phylogenetics And Networks for Generalized Epidemics in Africa) consortium has generated over 18 000 NGS HIV sequences from five countries in sub-Saharan Africa. Using phylogenetic methods, we will identify characteristics of individuals or groups, which are most likely to be at risk of infection or at risk of infecting others.

SUMMARY

Combining phylogenetics, phylodynamics and epidemiology will allow PANGEA to highlight where prevention efforts should be focussed to reduce the HIV epidemic most effectively. To maximise the public health benefit of the data, PANGEA offers accreditation to external researchers, allowing them to access the data and join the consortium. We also welcome submissions of other HIV sequences from sub-Saharan Africa to the database.

Phylodynamic Model Adequacy Using Posterior Predictive Simulations

Rapidly evolving pathogens, such as viruses and bacteria, accumulate genetic change at a similar timescale over which their epidemiological processes occur, such that, it is possible to make inferences about their infectious spread using phylogenetic time-trees. For this purpose it is necessary to choose a phylodynamic model. However, the resulting inferences are contingent on whether the model adequately describes key features of the data. Model adequacy methods allow formal rejection of a model if it cannot generate the main features of the data. We present TreeModelAdequacy, a package for the popular BEAST2 software that allows assessing the adequacy of phylodynamic models. We illustrate its utility by analyzing phylogenetic trees from two viral outbreaks of Ebola and H1N1 influenza. The main features of the Ebola data were adequately described by the coalescent exponential-growth model, whereas the H1N1 influenza data were best described by the birth–death susceptible-infected-recovered model.

Reconstructing foot-and-mouth disease outbreaks: a methods comparison of transmission network models

A number of transmission network models are available that combine genomic and epidemiological data to reconstruct networks of who infected whom during infectious disease outbreaks. For such models to reliably inform decision-making they must be transparently validated, robust, and capable of producing accurate predictions within the short data collection and inference timeframes typical of outbreak responses. A lack of transparent multi-model comparisons reduces confidence in the accuracy of transmission network model outputs, negatively impacting on their more widespread use as decision-support tools. We undertook a formal comparison of the performance of nine published transmission network models based on a set of foot-and-mouth disease outbreaks simulated in a previously free country, with corresponding simulated phylogenies and genomic samples from animals on infected premises. Of the transmission network models tested, Lau’s systematic Bayesian integration framework was found to be the most accurate for inferring the transmission network and timing of exposures, correctly identifying the source of 73% of the infected premises (with 91% accuracy for sources with model support >0.80). The Structured COalescent Transmission Tree Inference provided the most accurate inference of molecular clock rates. This validation study points to which models might be reliably used to reconstruct similar future outbreaks and how to interpret the outputs to inform control. Further research could involve extending the best-performing models to explicitly represent within-host diversity so they can handle next-generation sequencing data, incorporating additional animal and farm-level covariates and combining predictions using Ensemble methods and other approaches.

Designing & Conducting Trials To Reliably Evaluate HIV Prevention Interventions

While much has been achieved, much remains to be accomplished in the science of preventing the spread of HIV infection. Clinical trials that are properly designed, conducted and analyzed are of integral importance in the pursuit of reliable insights about HIV prevention. As we build on previous scientific breakthroughs, there will be an increasing need for clinical trials to be designed to efficiently achieve insights without compromising their reliability and generalizability. Key design features should continue to include: 1) the use of randomization and evidence-based controls, 2) specifying the use of intention-to-treat analyses to preserve the integrity of randomization and to increase interpretability of results, 3) obtaining direct assessments of effects on clinical endpoints such as the risk of HIV infection, 4) using either superiority designs or non-inferiority designs with rigorous non-inferiority margins, and 5) enhancing generalizability through the choice of a relative risk rather than risk difference metric. When interventions have complementary and potentially synergistic effects, factorial designs should be considered to increase efficiency as well as to obtain clinically important insights about interaction and the contribution of component interventions to the efficacy and safety of combination regimens. Key trial conduct issues include timely enrollment of participants at high HIV risk recruited from populations with high viral burden, obtaining 'best real-world achievable' levels of adherence to the interventions being assessed and ensuring high levels of retention. High quality of trial conduct occurs through active rather than passive monitoring, using pre-specified targeted levels of performance with defined methods to achieve those targets. During trial conduct, active monitoring of the performance standards not only holds the trial leaders accountable but also can assist in the development and implementation of creative alternative approaches to increase the quality of trial conduct. Designing, conducting and analyzing HIV prevention trials with the quality needed to obtain reliable insights is an ethical as well as scientific imperative.

Ethical considerations in global HIV phylogenetic research

Phylogenetic analysis of pathogens is an increasingly powerful way to reduce the spread of epidemics, including HIV. As a result, phylogenetic approaches are becoming embedded in public health and research programmes, as well as outbreak responses, presenting unique ethical, legal, and social issues that are not adequately addressed by existing bioethics literature. We formed a multidisciplinary working group to explore the ethical issues arising from the design of, conduct in, and use of results from HIV phylogenetic studies, and to propose recommendations to minimise the associated risks to both individuals and groups. We identified eight key ethical domains, within which we highlighted factors that make HIV phylogenetic research unique. In this Review, we endeavoured to provide a framework to assist researchers, public health practitioners, and funding institutions to ensure that HIV phylogenetic studies are designed, done, and disseminated in an ethical manner. Our conclusions also have broader relevance for pathogen phylogenetics.

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.

HIV-1 Transmission Clustering and Phylodynamics Highlight the Important Role of Young Men Who Have Sex with Men

More persons living with HIV reside in the Southern United States than in any other region, yet little is known about HIV molecular epidemiology in the South. We used cluster and phylodynamic analyses to evaluate HIV transmission patterns in middle Tennessee. We performed cross-sectional analyses of HIV-1 pol sequences and clinical data collected from 2001 to 2015 among persons attending the Vanderbilt Comprehensive Care Clinic. Transmission clusters were identified using maximum likelihood phylogenetics and patristic distance differences. Demographic, risk behavior, and clinical factors were assessed evaluating “active” clusters (clusters including sequences sampled 2011–2015) and associations estimated with logistic regression. Transmission risk ratios for men who have sex with men (MSM) were estimated with phylodynamic models. Among 2915 persons (96% subtype-B sequences), 963 (33%) were members of 292 clusters (distance ≤1.5%, size range 2–39). Most clusters (62%, n = 690 persons) were active, either being newly identified (n = 80) or showing expansion on existing clusters (n = 101). Correlates of active clustering among persons with sequences collected during 2011–2015 included MSM risk and ≤30 years of age. Active clusters were significantly more concentrated in MSM and younger persons than historical clusters. Young MSM (YMSM) (≤26.4 years) had high estimated transmission risk [risk ratio = 4.04 (2.85–5.65) relative to older MSM] and were much more likely to transmit to YMSM. In this Tennessee cohort, transmission clusters over time were more concentrated by MSM and younger age, with high transmission risk among and between YMSM, highlighting the importance of interventions among this group. Detecting active clusters could help direct interventions to disrupt ongoing transmission chains.

Inferring the age difference in HIV transmission pairs by applying phylogenetic methods on the HIV transmission network of the Swiss HIV Cohort Study

Age-mixing patterns are of key importance for understanding the dynamics of human immunodeficiency virus (HIV)-epidemics and target public health interventions. We use the densely sampled Swiss HIV Cohort Study (SHCS) resistance database to study the age difference at infection in HIV transmission pairs using phylogenetic methods. In addition, we investigate whether the mean age difference of pairs in the phylogenetic tree is influenced by sampling as well as by additional distance thresholds for including pairs. HIV-1 pol-sequences of 11,922 SHCS patients and approximately 240,000 Los Alamos background sequences were used to build a phylogenetic tree. Using this tree, 100 per cent down to 1 per cent of the tips were sampled repeatedly to generate pruned trees (N = 500 for each sample proportion), of which pairs of SHCS patients were extracted. The mean of the absolute age differences of the pairs, measured as the absolute difference of the birth years, was analyzed with respect to this sample proportion and a distance criterion for inclusion of the pairs. In addition, the transmission groups men having sex with men (MSM), intravenous drug users (IDU), and heterosexuals (HET) were analyzed separately. Considering the tree with all 11,922 SHCS patients, 2,991 pairs could be extracted, with 954 (31.9 per cent) MSM-pairs, 635 (21.2 per cent) HET-pairs, 414 (13.8 per cent) IDU-pairs, and 352 (11.8 per cent) HET/IDU-pairs. For all transmission groups, the age difference at infection was significantly (P < 0.001) smaller for pairs in the tree compared with randomly assigned pairs, meaning that patients of similar age are more likely to be pairs. The mean age difference in the phylogenetic analysis, using a fixed distance of 0.05, was 9.2, 9.0, 7.3 and 5.6 years for MSM-, HET-, HET/IDU-, and IDU-pairs, respectively. Decreasing the cophenetic distance threshold from 0.05 to 0.01 significantly decreased the mean age difference. Similarly, repeated sampling of 100 per cent down to 1 per cent of the tips revealed an increased age difference at lower sample proportions. HIV-transmission is age-assortative, but the age difference of transmission pairs detected by phylogenetic analyses depends on both sampling proportion and distance criterion. The mean age difference decreases when using more conservative distance thresholds, implying an underestimation of age-assortativity when using liberal distance criteria. Similarly, overestimation of the mean age difference occurs for pairs from sparsely sampled trees, as it is often the case in sub-Saharan Africa.

HIV-1 diversity among young women in rural South Africa: HPTN 068

Background

South Africa has one of the highest rates of HIV-1 (HIV) infection world-wide, with the highest rates among young women. We analyzed the molecular epidemiology and evolutionary history of HIV in young women attending high school in rural South Africa.

Methods

Samples were obtained from the HPTN 068 randomized controlled trial, which evaluated the effect of cash transfers for school attendance on HIV incidence in women aged 13–20 years (Mpumalanga province, 2011–2015). Plasma samples from HIV-infected participants were analyzed using the ViroSeq HIV-1 Genotyping assay. Phylogenetic analysis was performed using 200 pol gene study sequences and 2,294 subtype C reference sequences from South Africa. Transmission clusters were identified using Cluster Picker and HIV-TRACE, and were characterized using demographic and other epidemiological data. Phylodynamic analyses were performed using the BEAST software.

Results

The study enrolled 2,533 young women who were followed through their expected high school graduation date (main study); some participants had a post-study assessment (follow-up study). Two-hundred-twelve of 2,533 enrolled young women had HIV infection. HIV pol sequences were obtained for 94% (n = 201/212) of the HIV-infected participants. All but one of the sequences were HIV-1 subtype C; the non-C subtype sequence was excluded from further analysis. Median pairwise genetic distance between the subtype C sequences was 6.4% (IQR: 5.6–7.2). Overall, 26% of study sequences fell into 21 phylogenetic clusters with 2–6 women per cluster. Thirteen (62%) clusters included women who were HIV-infected at enrollment. Clustering was not associated with study arm, demographic or other epidemiological factors. The estimated date of origin of HIV subtype C in the study population was 1958 (95% highest posterior density [HPD]: 1931–1980), and the median estimated substitution rate among study pol sequences was 1.98x10^-3 (95% HPD: 1.15x10^-3–2.81x10^-3) per site per year.

Conclusions

Phylogenetic analysis suggests that multiple HIV subtype C sublineages circulate among school age girls in South Africa. There were no substantive differences in the molecular epidemiology of HIV between control and intervention arms in the HPTN 068 trial.

Tracking external introductions of HIV using phylodynamics reveals a major source of infections in rural KwaZulu-Natal, South Africa

Despite increasing access to antiretrovirals, HIV incidence in rural KwaZulu-Natal remains among the highest ever reported in Africa. While many epidemiological factors have been invoked to explain such high incidence, widespread human mobility and viral movement suggest that transmission between communities may be a major source of new infections. High cross-community transmission rates call into question how effective increasing the coverage of antiretroviral therapy locally will be at preventing new infections, especially if many new cases arise from external introductions. To help address this question, we use a phylodynamic model to reconstruct epidemic dynamics and estimate the relative contribution of local transmission versus external introductions to overall incidence in KwaZulu-Natal from HIV-1 phylogenies. By comparing our results with population-based surveillance data, we show that we can reliably estimate incidence from viral phylogenies once viral movement in and out of the local population is accounted for. Our analysis reveals that early epidemic dynamics were largely driven by external introductions. More recently, we estimate that 35 per cent (95% confidence interval: 20-60%) of new infections arise from external introductions. These results highlight the growing need to consider larger-scale regional transmission dynamics when designing and testing prevention strategies.

Genotypic and Phylogenetic Insights on Prevention of the Spread of HIV-1 and Drug Resistance in "Real-World" Settings

HIV continues to spread among vulnerable heterosexual (HET), Men-having-Sex with Men (MSM) and intravenous drug user (IDU) populations, influenced by a complex array of biological, behavioral and societal factors. Phylogenetics analyses of large sequence datasets from national drug resistance testing programs reveal the evolutionary interrelationships of viral strains implicated in the dynamic spread of HIV in different regional settings. Viral phylogenetics can be combined with demographic and behavioral information to gain insights on epidemiological processes shaping transmission networks at the population-level. Drug resistance testing programs also reveal emergent mutational pathways leading to resistance to the 23 antiretroviral drugs used in HIV-1 management in low-, middle- and high-income settings. This article describes how genotypic and phylogenetic information from Quebec and elsewhere provide critical information on HIV transmission and resistance, Cumulative findings can be used to optimize public health strategies to tackle the challenges of HIV in “real-world” settings.

HIV-1 full-genome phylogenetics of generalized epidemics in sub-Saharan Africa: impact of missing nucleotide characters in next-generation sequences

To characterize HIV-1 transmission dynamics in regions where the burden of HIV-1 is greatest, the “Phylogenetics and Networks for Generalised HIV Epidemics in Africa” consortium (PANGEA-HIV) is sequencing full-genome viral isolates from across sub-Saharan Africa. We report the first 3,985 PANGEA-HIV consensus sequences from four cohort sites (Rakai Community Cohort Study, n = 2,833; MRC/UVRI Uganda, n = 701; Mochudi Prevention Project, n = 359; Africa Health Research Institute Resistance Cohort, n = 92). Next-generation sequencing success rates varied: more than 80% of the viral genome from the gag to the nef genes could be determined for all sequences from South Africa, 75% of sequences from Mochudi, 60% of sequences from MRC/UVRI Uganda, and 22% of sequences from Rakai. Partial sequencing failure was primarily associated with low viral load, increased for amplicons closer to the 3′ end of the genome, was not associated with subtype diversity except HIV-1 subtype D, and remained significantly associated with sampling location after controlling for other factors. We assessed the impact of the missing data patterns in PANGEA-HIV sequences on phylogeny reconstruction in simulations. We found a threshold in terms of taxon sampling below which the patchy distribution of missing characters in next-generation sequences (NGS) has an excess negative impact on the accuracy of HIV-1 phylogeny reconstruction, which is attributable to tree reconstruction artifacts that accumulate when branches in viral trees are long. The large number of PANGEA-HIV sequences provides unprecedented opportunities for evaluating HIV-1 transmission dynamics across sub-Saharan Africa and identifying prevention opportunities. Molecular epidemiological analyses of these data must proceed cautiously because sequence sampling remains below the identified threshold and a considerable negative impact of missing characters on phylogeny reconstruction is expected.

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.

Inferring epidemiological parameters from phylogenies using regression-ABC: A comparative study

Inferring epidemiological parameters such as the R₀ from time-scaled phylogenies is a timely challenge. Most current approaches rely on likelihood functions, which raise specific issues that range from computing these functions to finding their maxima numerically. Here, we present a new regression-based Approximate Bayesian Computation (ABC) approach, which we base on a large variety of summary statistics intended to capture the information contained in the phylogeny and its corresponding lineage-through-time plot. The regression step involves the Least Absolute Shrinkage and Selection Operator (LASSO) method, which is a robust machine learning technique. It allows us to readily deal with the large number of summary statistics, while avoiding resorting to Markov Chain Monte Carlo (MCMC) techniques. To compare our approach to existing ones, we simulated target trees under a variety of epidemiological models and settings, and inferred parameters of interest using the same priors. We found that, for large phylogenies, the accuracy of our regression-ABC is comparable to that of likelihood-based approaches involving birth-death processes implemented in BEAST2. Our approach even outperformed these when inferring the host population size with a Susceptible-Infected-Removed epidemiological model. It also clearly outperformed a recent kernel-ABC approach when assuming a Susceptible-Infected epidemiological model with two host types. Lastly, by re-analyzing data from the early stages of the recent Ebola epidemic in Sierra Leone, we showed that regression-ABC provides more realistic estimates for the duration parameters (latency and infectiousness) than the likelihood-based method. Overall, ABC based on a large variety of summary statistics and a regression method able to perform variable selection and avoid overfitting is a promising approach to analyze large phylogenies.

Given the rapid evolution of many pathogens, analysing their genomes by means of phylogenies can inform us about how they spread. This is the focus of the field known as “phylodynamics”. Most existing methods inferring epidemiological parameters from virus phylogenies are limited by the difficulty of handling complex likelihood functions, which commonly incorporate latent variables. Here, we use an alternative method known as regression-based Approximate Bayesian Computation (ABC), which circumvents this problem by using simulations and dataset comparisons. Since phylogenies are difficult to compare to one another, we introduce many summary statistics to describe them and take advantage of current machine learning techniques able to perform variable selection. We show that the accuracy we reach is comparable to that of existing methods. This accuracy increases with phylogeny size and can even be higher than that of existing methods for some parameters. Overall, regression-based ABC opens new perspectives to infer epidemiological parameters from large phylogenies.

Using nearly full-genome HIV sequence data improves phylogeny reconstruction in a simulated epidemic

HIV molecular epidemiology studies analyse viral pol gene sequences due to their availability, but whole genome sequencing allows to use other genes. We aimed to determine what gene(s) provide(s) the best approximation to the real phylogeny by analysing a simulated epidemic (created as part of the PANGEA_HIV project) with a known transmission tree. We sub-sampled a simulated dataset of 4662 sequences into different combinations of genes (gag-pol-env, gag-pol, gag, pol, env and partial pol) and sampling depths (100%, 60%, 20% and 5%), generating 100 replicates for each case. We built maximum-likelihood trees for each combination using RAxML (GTR + Γ), and compared their topologies to the corresponding true tree’s using CompareTree. The accuracy of the trees was significantly proportional to the length of the sequences used, with the gag-pol-env datasets showing the best performance and gag and partial pol sequences showing the worst. The lowest sampling depths (20% and 5%) greatly reduced the accuracy of tree reconstruction and showed high variability among replicates, especially when using the shortest gene datasets. In conclusion, using longer sequences derived from nearly whole genomes will improve the reliability of phylogenetic reconstruction. With low sample coverage, results can be highly variable, particularly when based on short sequences.