A Surrogate Function for One-Dimensional Phylogenetic Likelihoods.

Phylogenetics has seen a steady increase in data set size and substitution model complexity, which require increasing amounts of computational power to compute likelihoods. This motivates strategies to approximate the likelihood functions for branch length optimization and Bayesian sampling. In this article, we develop an approximation to the 1D likelihood function as parametrized by a single branch length. Our method uses a four-parameter surrogate function abstracted from the simplest phylogenetic likelihood function, the binary symmetric model. We show that it offers a surrogate that can be fit over a variety of branch lengths, that it is applicable to a wide variety of models and trees, and that it can be used effectively as a proposal mechanism for Bayesian sampling. The method is implemented as a stand-alone open-source C library for calling from phylogenetics algorithms; it has proven essential for good performance of our online phylogenetic algorithm sts.

Effective Online Bayesian Phylogenetics via Sequential Monte Carlo with Guided Proposals.

Modern infectious disease outbreak surveillance produces continuous streams of sequence data which require phylogenetic analysis as data arrives. Current software packages for Bayesian phylogenetic inference are unable to quickly incorporate new sequences as they become available, making them less useful for dynamically unfolding evolutionary stories. This limitation can be addressed by applying a class of Bayesian statistical inference algorithms called sequential Monte Carlo (SMC) to conduct online inference, wherein new data can be continuously incorporated to update the estimate of the posterior probability distribution. In this article, we describe and evaluate several different online phylogenetic sequential Monte Carlo (OPSMC) algorithms. We show that proposing new phylogenies with a density similar to the Bayesian prior suffers from poor performance, and we develop "guided" proposals that better match the proposal density to the posterior. Furthermore, we show that the simplest guided proposals can exhibit pathological behavior in some situations, leading to poor results, and that the situation can be resolved by heating the proposal density. The results demonstrate that relative to the widely used MCMC-based algorithm implemented in MrBayes, the total time required to compute a series of phylogenetic posteriors as sequences arrive can be significantly reduced by the use of OPSMC, without incurring a significant loss in accuracy.

Comprehensive sieve analysis of breakthrough HIV-1 sequences in the RV144 vaccine efficacy trial.

The RV144 clinical trial showed the partial efficacy of a vaccine regimen with an estimated vaccine efficacy (VE) of 31% for protecting low-risk Thai volunteers against acquisition of HIV-1. The impact of vaccine-induced immune responses can be investigated through sieve analysis of HIV-1 breakthrough infections (infected vaccine and placebo recipients). A V1/V2-targeted comparison of the genomes of HIV-1 breakthrough viruses identified two V2 amino acid sites that differed between the vaccine and placebo groups. Here we extended the V1/V2 analysis to the entire HIV-1 genome using an array of methods based on individual sites, k-mers and genes/proteins. We identified 56 amino acid sites or "signatures" and 119 k-mers that differed between the vaccine and placebo groups. Of those, 19 sites and 38 k-mers were located in the regions comprising the RV144 vaccine (Env-gp120, Gag, and Pro). The nine signature sites in Env-gp120 were significantly enriched for known antibody-associated sites (p = 0.0021). In particular, site 317 in the third variable loop (V3) overlapped with a hotspot of antibody recognition, and sites 369 and 424 were linked to CD4 binding site neutralization. The identified signature sites significantly covaried with other sites across the genome (mean = 32.1) more than did non-signature sites (mean = 0.9) (p < 0.0001), suggesting functional and/or structural relevance of the signature sites. Since signature sites were not preferentially restricted to the vaccine immunogens and because most of the associations were insignificant following correction for multiple testing, we predict that few of the genetic differences are strongly linked to the RV144 vaccine-induced immune pressure. In addition to presenting results of the first complete-genome analysis of the breakthrough infections in the RV144 trial, this work describes a set of statistical methods and tools applicable to analysis of breakthrough infection genomes in general vaccine efficacy trials for diverse pathogens.

Risk of drug resistance among persons acquiring HIV within a randomized clinical trial of single- or dual-agent preexposure prophylaxis.

BACKGROUND: Preexposure prophylaxis (PrEP) with emtricitabine plus tenofovir disoproxil fumarate (FTC/TDF) or TDF alone reduces the risk of human immunodeficiency virus (HIV) acquisition. Understanding the risk of antiretroviral resistance selected by PrEP during breakthrough infections is important because of the risk of treatment failure during subsequent antiretroviral use. METHODS: Within the largest randomized trial of FTC/TDF versus TDF as PrEP, plasma samples were tested for HIV with resistance mutations associated with FTC (K65R and M184IV) and TDF (K65R and K70E), using 454 sequencing. RESULTS: Of 121 HIV seroconverters, 25 received FTC/TDF, 38 received TDF, and 58 received placebo. Plasma drug levels in 26 individuals indicated PrEP use during or after HIV acquisition, of which 5 had virus with resistance mutations associated with their PrEP regimen. Among those with PrEP drug detected during infection, resistance was more frequent in the FTC/TDF arm (4 of 7 [57%]), compared with the TDF arm (1 of 19 [5.3%]; P = .01), owing to the FTC-associated mutation M184IV. Of these cases, 3 had unrecognized acute infection at PrEP randomization, and 2 were HIV negative at enrollment. CONCLUSIONS: These results suggest that resistance selected by PrEP is rare but can occur both with PrEP initiation during acute seronegative HIV infection and in PrEP breakthrough infections and that FTC is associated with a greater frequency of resistance mutations than TDF.

HIV-1 superinfection occurs less frequently than initial infection in a cohort of high-risk Kenyan women.

HIV superinfection (reinfection) has been reported in several settings, but no study has been designed and powered to rigorously compare its incidence to that of initial infection. Determining whether HIV infection reduces the risk of superinfection is critical to understanding whether an immune response to natural HIV infection is protective. This study compares the incidence of initial infection and superinfection in a prospective seroincident cohort of high-risk women in Mombasa, Kenya. A next-generation sequencing-based pipeline was developed to screen 129 women for superinfection. Longitudinal plasma samples at <6 months, >2 years and one intervening time after initial HIV infection were analyzed. Amplicons in three genome regions were sequenced and a median of 901 sequences obtained per gene per timepoint. Phylogenetic evidence of polyphyly, confirmed by pairwise distance analysis, defined superinfection. Superinfection timing was determined by sequencing virus from intervening timepoints. These data were combined with published data from 17 additional women in the same cohort, totaling 146 women screened. Twenty-one cases of superinfection were identified for an estimated incidence rate of 2.61 per 100 person-years (pys). The incidence rate of initial infection among 1910 women in the same cohort was 5.75 per 100 pys. Andersen-Gill proportional hazards models were used to compare incidences, adjusting for covariates known to influence HIV susceptibility in this cohort. Superinfection incidence was significantly lower than initial infection incidence, with a hazard ratio of 0.47 (CI 0.29-0.75, p = 0.0019). This lower incidence of superinfection was only observed >6 months after initial infection. This is the first adequately powered study to report that HIV infection reduces the risk of reinfection, raising the possibility that immune responses to natural infection are partially protective. The observation that superinfection risk changes with time implies a window of protection that coincides with the maturation of HIV-specific immunity.

Phylogenomic analysis reveals dynamic evolutionary history of the Drosophila heterochromatin protein 1 (HP1) gene family.

Heterochromatin is the gene-poor, satellite-rich eukaryotic genome compartment that supports many essential cellular processes. The functional diversity of proteins that bind and often epigenetically define heterochromatic DNA sequence reflects the diverse functions supported by this enigmatic genome compartment. Moreover, heterogeneous signatures of selection at chromosomal proteins often mirror the heterogeneity of evolutionary forces that act on heterochromatic DNA. To identify new such surrogates for dissecting heterochromatin function and evolution, we conducted a comprehensive phylogenomic analysis of the Heterochromatin Protein 1 gene family across 40 million years of Drosophila evolution. Our study expands this gene family from 5 genes to at least 26 genes, including several uncharacterized genes in Drosophila melanogaster. The 21 newly defined HP1s introduce unprecedented structural diversity, lineage-restriction, and germline-biased expression patterns into the HP1 family. We find little evidence of positive selection at these HP1 genes in both population genetic and molecular evolution analyses. Instead, we find that dynamic evolution occurs via prolific gene gains and losses. Despite this dynamic gene turnover, the number of HP1 genes is relatively constant across species. We propose that karyotype evolution drives at least some HP1 gene turnover. For example, the loss of the male germline-restricted HP1E in the obscura group coincides with one episode of dramatic karyotypic evolution, including the gain of a neo-Y in this lineage. This expanded compendium of ovary- and testis-restricted HP1 genes revealed by our study, together with correlated gain/loss dynamics and chromosome fission/fusion events, will guide functional analyses of novel roles supported by germline chromatin.

Nestly--a framework for running software with nested parameter choices and aggregating results.

UNLABELLED: The execution of a software application or pipeline using various combinations of parameters and inputs is a common task in bioinformatics. In the absence of a specialized tool to organize, streamline and formalize this process, scientists must write frequently complex scripts to perform these tasks. We present nestly, a Python package to facilitate running tools with nested combinations of parameters and inputs. nestly provides three components. First, a module to build nested directory structures corresponding to choices of parameters. Second, the nestrun script to run a given command using each set of parameter choices. Third, the nestagg script to aggregate results of the individual runs into a CSV file, as well as support for more complex aggregation. We also include a module for easily specifying nested dependencies for the SCons build tool, enabling incremental builds. AVAILABILITY: Source, documentation and tutorial examples are available at http://github.com/fhcrc/nestly. nestly can be installed from the Python Package Index via pip; it is open source (MIT license).

Minimizing the average distance to a closest leaf in a phylogenetic tree.

When performing an analysis on a collection of molecular sequences, it can be convenient to reduce the number of sequences under consideration while maintaining some characteristic of a larger collection of sequences. For example, one may wish to select a subset of high-quality sequences that represent the diversity of a larger collection of sequences. One may also wish to specialize a large database of characterized "reference sequences" to a smaller subset that is as close as possible on average to a collection of "query sequences" of interest. Such a representative subset can be useful whenever one wishes to find a set of reference sequences that is appropriate to use for comparative analysis of environmentally derived sequences, such as for selecting "reference tree" sequences for phylogenetic placement of metagenomic reads. In this article, we formalize these problems in terms of the minimization of the Average Distance to the Closest Leaf (ADCL) and investigate algorithms to perform the relevant minimization. We show that the greedy algorithm is not effective, show that a variant of the Partitioning Around Medoids (PAM) heuristic gets stuck in local minima, and develop an exact dynamic programming approach. Using this exact program we note that the performance of PAM appears to be good for simulated trees, and is faster than the exact algorithm for small trees. On the other hand, the exact program gives solutions for all numbers of leaves less than or equal to the given desired number of leaves, whereas PAM only gives a solution for the prespecified number of leaves. Via application to real data, we show that the ADCL criterion chooses chimeric sequences less often than random subsets, whereas the maximization of phylogenetic diversity chooses them more often than random. These algorithms have been implemented in publicly available software.

Preexposure prophylaxis-selected drug resistance decays rapidly after drug cessation.

OBJECTIVE: Resistance to emtricitabine plus tenofovir disoproxil fumarate (FTC/TDF) or TDF alone used as preexposure prophylaxis (PrEP) has been detected in individuals who initiated PrEP during unrecognized acute HIV infection and, rarely, in PrEP breakthrough infections. PrEP-selected resistance could alter future treatment options, and therefore we sought to determine how long resistance persisted after PrEP cessation. METHODS: The Partners PrEP Study was a randomized placebo-controlled trial of FTC/TDF or TDF as PrEP for HIV prevention. We previously reported that PrEP-related mutations (K65R, K70E or M184IV) were detected by 454 sequencing following seroconversion in nine individuals who acquired HIV during the Partners PrEP Study. In the current study, we used 454 sequencing to detect and quantify PrEP-related mutations in HIV RNA-positive plasma samples prior to seroconversion, as well as in plasma from 6, 12, and 24 months after PrEP cessation from these nine individuals. RESULTS: HIV RNA-positive, antibody-negative samples were available prior to seroconversion for four of nine individuals with resistance detected at seroconversion. In all four cases, K65R, K70E and M184IV were not detected prior to seroconversion, suggesting PrEP-related resistance was selected and not transmitted. All PrEP-selected mutations were no longer detectable by 6 months after PrEP cessation and remained undetectable at 12 and 24 months in the absence of antiretroviral therapy. CONCLUSION: Using highly sensitive assays, PrEP-selected resistance in plasma decays below detection by 6 months following drug cessation and remains undetectable for at least 24 months. Even high levels of resistance mutations during acute infection decay rapidly in the absence of ongoing PrEP exposure.

Quantifying evolutionary constraints on B-cell affinity maturation.

The antibody repertoire of each individual is continuously updated by the evolutionary process of B-cell receptor (BCR) mutation and selection. It has recently become possible to gain detailed information concerning this process through high-throughput sequencing. Here, we develop modern statistical molecular evolution methods for the analysis of B-cell sequence data, and then apply them to a very deep short-read dataset of BCRs. We find that the substitution process is conserved across individuals but varies significantly across gene segments. We investigate selection on BCRs using a novel method that side-steps the difficulties encountered by previous work in differentiating between selection and motif-driven mutation; this is done through stochastic mapping and empirical Bayes estimators that compare the evolution of in-frame and out-of-frame rearrangements. We use this new method to derive a per-residue map of selection, which provides a more nuanced view of the constraints on framework and variable regions.

Bacterial Composition of the Human Upper Gastrointestinal Tract Microbiome Is Dynamic and Associated with Genomic Instability in a Barrett's Esophagus Cohort.

BACKGROUND: The incidence of esophageal adenocarcinoma (EAC) has increased nearly five-fold over the last four decades in the United States. Barrett's esophagus, the replacement of the normal squamous epithelial lining with a mucus-secreting columnar epithelium, is the only known precursor to EAC. Like other parts of the gastrointestinal (GI) tract, the esophagus hosts a variety of bacteria and comparisons among published studies suggest bacterial communities in the stomach and esophagus differ. Chronic infection with Helicobacter pylori in the stomach has been inversely associated with development of EAC, but the mechanisms underlying this association remain unclear. METHODOLOGY: The bacterial composition in the upper GI tract was characterized in a subset of participants (n=12) of the Seattle Barrett's Esophagus Research cohort using broad-range 16S PCR and pyrosequencing of biopsy and brush samples collected from squamous esophagus, Barrett's esophagus, stomach corpus and stomach antrum. Three of the individuals were sampled at two separate time points. Prevalence of H. pylori infection and subsequent development of aneuploidy (n=339) and EAC (n=433) was examined in a larger subset of this cohort. RESULTS/SIGNIFICANCE: Within individuals, bacterial communities of the stomach and esophagus showed overlapping community membership. Despite closer proximity, the stomach antrum and corpus communities were less similar than the antrum and esophageal samples. Re-sampling of study participants revealed similar upper GI community membership in two of three cases. In this Barrett's esophagus cohort, Streptococcus and Prevotella species dominate the upper GI and the ratio of these two species is associated with waist-to-hip ratio and hiatal hernia length, two known EAC risk factors in Barrett's esophagus. H. pylori-positive individuals had a significantly decreased incidence of aneuploidy and a non-significant trend toward lower incidence of EAC.

Abundance-weighted phylogenetic diversity measures distinguish microbial community states and are robust to sampling depth.

In microbial ecology studies, the most commonly used ways of investigating alpha (within-sample) diversity are either to apply non-phylogenetic measures such as Simpson's index to Operational Taxonomic Unit (OTU) groupings, or to use classical phylogenetic diversity (PD), which is not abundance-weighted. Although alpha diversity measures that use abundance information in a phylogenetic framework do exist, they are not widely used within the microbial ecology community. The performance of abundance-weighted phylogenetic diversity measures compared to classical discrete measures has not been explored, and the behavior of these measures under rarefaction (sub-sampling) is not yet clear. In this paper we compare the ability of various alpha diversity measures to distinguish between different community states in the human microbiome for three different datasets. We also present and compare a novel one-parameter family of alpha diversity measures, BWPDθ, that interpolates between classical phylogenetic diversity (PD) and an abundance-weighted extension of PD. Additionally, we examine the sensitivity of these phylogenetic diversity measures to sampling, via computational experiments and by deriving a closed form solution for the expectation of phylogenetic quadratic entropy under re-sampling. On the three datasets, a phylogenetic measure always performed best, and two abundance-weighted phylogenetic diversity measures were the only measures ranking in the top four across all datasets. OTU-based measures, on the other hand, are less effective in distinguishing community types. In addition, abundance-weighted phylogenetic diversity measures are less sensitive to differing sampling intensity than their unweighted counterparts. Based on these results we encourage the use of abundance-weighted phylogenetic diversity measures, especially for cases such as microbial ecology where species delimitation is difficult.

Rapid 16S rRNA next-generation sequencing of polymicrobial clinical samples for diagnosis of complex bacterial infections.

Classifying individual bacterial species comprising complex, polymicrobial patient specimens remains a challenge for culture-based and molecular microbiology techniques in common clinical use. We therefore adapted practices from metagenomics research to rapidly catalog the bacterial composition of clinical specimens directly from patients, without need for prior culture. We have combined a semiconductor deep sequencing protocol that produces reads spanning 16S ribosomal RNA gene variable regions 1 and 2 (∼360 bp) with a de-noising pipeline that significantly improves the fraction of error-free sequences. The resulting sequences can be used to perform accurate genus- or species-level taxonomic assignment. We explore the microbial composition of challenging, heterogeneous clinical specimens by deep sequencing, culture-based strain typing, and Sanger sequencing of bulk PCR product. We report that deep sequencing can catalog bacterial species in mixed specimens from which usable data cannot be obtained by conventional clinical methods. Deep sequencing a collection of sputum samples from cystic fibrosis (CF) patients reveals well-described CF pathogens in specimens where they were not detected by standard clinical culture methods, especially for low-prevalence or fastidious bacteria. We also found that sputa submitted for CF diagnostic workup can be divided into a limited number of groups based on the phylogenetic composition of the airway microbiota, suggesting that metagenomic profiling may prove useful as a clinical diagnostic strategy in the future. The described method is sufficiently rapid (theoretically compatible with same-day turnaround times) and inexpensive for routine clinical use.

The ability of primate lentiviruses to degrade the monocyte restriction factor SAMHD1 preceded the birth of the viral accessory protein Vpx.

The human SAMHD1 protein potently restricts lentiviral infection in dendritic cells and monocyte/macrophages but is antagonized by the primate lentiviral protein Vpx, which targets SAMHD1 for degradation. However, only two of eight primate lentivirus lineages encode Vpx, whereas its paralog, Vpr, is conserved across all extant primate lentiviruses. We find that not only multiple Vpx but also some Vpr proteins are able to degrade SAMHD1, and such antagonism led to dramatic positive selection of SAMHD1 in the primate subfamily Cercopithecinae. Residues that have evolved under positive selection precisely determine sensitivity to Vpx/Vpr degradation and alter binding specificity. By overlaying these functional analyses on a phylogenetic framework of Vpr and Vpx evolution, we can decipher the chronology of acquisition of SAMHD1-degrading abilities in lentiviruses. We conclude that vpr neofunctionalized to degrade SAMHD1 even prior to the birth of a separate vpx gene, thereby initiating an evolutionary arms race with SAMHD1.

Low-frequency nevirapine resistance at multiple sites may predict treatment failure in infants on nevirapine-based treatment.

BACKGROUND: Resistance commonly arises in infants exposed to single-dose nevirapine (sdNVP) for prevention of mother to child transmission. Although K103N and Y181C are common following sdNVP, multiple other mutations also confer NVP resistance. It remains unclear whether specific NVP resistance mutations or combinations of mutations predict virologic failure in infants when present at low frequencies before NVP-based treatment. METHODS: Twenty sdNVP-exposed infants who were subsequently treated with NVP-based highly active antiretroviral therapy were examined. Pretreatment plasma samples were tested for the presence of NVP resistance mutations by allele-specific polymerase chain reaction for K103N and Y181C and ultradeep pyrosequencing (UDPS) for all primary NVP mutations. Viral levels were determined every 3 months for up to 24 months on NVP-highly active antiretroviral therapy. Cox proportional hazard models were used to determine correlates of viral failure. RESULTS: The NVP resistance mutations K103N or Y181C were detected in pretreatment plasma samples in 6 infants by allele-specific polymerase chain reaction. NVP resistance at these or other sites was detectable by UDPS in 10 of 20 infants tested. Virologic failure occurred in 50% of infants with any NVP resistance mutations detected, whereas only 20% of infants without resistance experienced viral failure, but the difference was not significant (P = 0.19). An increase in the number of NVP resistance mutations detectable by UDPS in an infant was significantly associated with an increased risk of virologic failure [HR = 1.79 (95% confidence interval: 1.07 to 2.99), P = 0.027]. CONCLUSIONS: Low frequencies of multiple NVP resistance mutations, in addition to K103N and Y181C, present in infants before NVP-based treatment may predict treatment outcome.

Bacterial communities in women with bacterial vaginosis: high resolution phylogenetic analyses reveal relationships of microbiota to clinical criteria.

BACKGROUND: Bacterial vaginosis (BV) is a common condition that is associated with numerous adverse health outcomes and is characterized by poorly understood changes in the vaginal microbiota. We sought to describe the composition and diversity of the vaginal bacterial biota in women with BV using deep sequencing of the 16S rRNA gene coupled with species-level taxonomic identification. We investigated the associations between the presence of individual bacterial species and clinical diagnostic characteristics of BV. METHODOLOGY/PRINCIPAL FINDINGS: Broad-range 16S rRNA gene PCR and pyrosequencing were performed on vaginal swabs from 220 women with and without BV. BV was assessed by Amsel's clinical criteria and confirmed by Gram stain. Taxonomic classification was performed using phylogenetic placement tools that assigned 99% of query sequence reads to the species level. Women with BV had heterogeneous vaginal bacterial communities that were usually not dominated by a single taxon. In the absence of BV, vaginal bacterial communities were dominated by either Lactobacillus crispatus or Lactobacillus iners. Leptotrichia amnionii and Eggerthella sp. were the only two BV-associated bacteria (BVABs) significantly associated with each of the four Amsel's criteria. Co-occurrence analysis revealed the presence of several sub-groups of BVABs suggesting metabolic co-dependencies. Greater abundance of several BVABs was observed in Black women without BV. CONCLUSIONS/SIGNIFICANCE: The human vaginal bacterial biota is heterogeneous and marked by greater species richness and diversity in women with BV; no species is universally present. Different bacterial species have different associations with the four clinical criteria, which may account for discrepancies often observed between Amsel and Nugent (Gram stain) diagnostic criteria. Several BVABs exhibited race-dependent prevalence when analyzed in separate groups by BV status which may contribute to increased incidence of BV in Black women. Tools developed in this project can be used to study microbial ecology in diverse settings at high resolution.