Search

Clustering infections by genetic similarity is a popular technique for identifying potential outbreaks of infectious disease, in part because sequences are now routinely collected for clinical management of many infections. A diverse number of nonparametric clustering methods have been developed for this purpose. These methods are generally intuitive, rapid to compute, and readily scale with large data sets. However, we have found that nonparametric clustering methods can be biased towards identifying clusters of diagnosis—where individuals are sampled sooner post-infection—rather than the clusters of rapid transmission that are meant to be potential foci for public health efforts. We develop a fundamentally new approach to genetic clustering based on fitting a Markov-modulated Poisson process (MMPP), which represents the evolution of transmission rates along the tree relating different infections. We evaluated this model-based method alongside five nonparametric clustering methods using both simulated and actual HIV sequence data sets. For simulated clusters of rapid transmission, the MMPP clustering method obtained higher mean sensitivity (85%) and specificity (91%) than the nonparametric methods. When we applied these clustering methods to published sequences from a study of HIV-1 genetic clusters in Seattle, USA, we found that the MMPP method categorized about half (46%) as many individuals to clusters compared to the other methods. Furthermore, the mean internal branch lengths that approximate transmission rates were significantly shorter in clusters extracted using MMPP, but not by other methods. We determined that the computing time for the MMPP method scaled linearly with the size of trees, requiring about 30 seconds for a tree of 1,000 tips and about 20 minutes for 50,000 tips on a single computer. This new approach to genetic clustering has significant implications for the application of pathogen sequence analysis to public health, where it is critical to robustly and accurately identify clusters for the most cost-effective deployment of outbreak management and prevention resources., Author summary Many pathogens evolve so rapidly that they accumulate genetic differences within a host before becoming transmitted to the next host. Consequently, clusters of sampled infections with nearly identical genomes may reveal outbreaks of recent or ongoing transmissions. There is rapidly growing interest in using model-free genetic clustering methods to guide public health responses to epidemics in near real-time, including HIV, Ebola virus and tuberculosis. However, we show that current methods are relatively ineffective at detecting transmission outbreaks; instead, they are predominantly influenced by how infections are sampled from the population. We describe a fundamentally new approach to genetic clustering that is based on modelling changes in transmission rates during the spread of the epidemic. We use simulated and real pathogen sequence data sets to demonstrate that this model-based approach is substantially more effective for detecting transmission outbreaks, and remains fast enough for real-time applications to large sequence databases.

A population of human immunodeficiency virus (HIV) within a host often descends from a single transmitted/founder virus. The high mutation rate of HIV, coupled with long delays between infection and diagnosis, make isolating and characterizing this strain a challenge. In theory, ancestral reconstruction could be used to recover this strain from sequences sampled in chronic infection; however, the accuracy of phylogenetic techniques in this context is unknown. To evaluate the accuracy of these methods, we applied ancestral reconstruction to a large panel of published longitudinal clonal and/or single-genome-amplification HIV sequence data sets with at least one intrapatient sequence set sampled within 6 months of infection or seroconversion ( n = 19,486 sequences, median [interquartile range] = 49 [20 to 86] sequences/set). The consensus of the earliest sequences was used as the best possible estimate of the transmitted/founder. These sequences were compared to ancestral reconstructions from sequences sampled at later time points using both phylogenetic and phylogeny-naive methods. Overall, phylogenetic methods conferred a 16% improvement in reproducing the consensus of early sequences, compared to phylogeny-naive methods. This relative advantage increased with intrapatient sequence diversity ( P < 10 −5 ) and the time elapsed between the earliest and subsequent samples ( P < 10 −5 ). However, neither approach performed well for reconstructing ancestral indel variation, especially within indel-rich regions of the HIV genome. Although further improvements are needed, our results indicate that phylogenetic methods for ancestral reconstruction significantly outperform phylogeny-naive alternatives, and we identify experimental conditions and study designs that can enhance accuracy of transmitted/founder virus reconstruction. IMPORTANCE When HIV is transmitted into a new host, most of the viruses fail to infect host cells. Consequently, an HIV infection tends to be descended from a single “founder” virus. A priority target for the vaccine research, these transmitted/founder viruses are difficult to isolate since newly infected individuals are often unaware of their status for months or years, by which time the virus population has evolved substantially. Here, we report on the potential use of evolutionary methods to reconstruct the genetic sequence of the transmitted/founder virus from its descendants at later stages of an infection. These methods can recover this ancestral sequence with an overall error rate of about 2.3%—about 15% more information than if we had ignored the evolutionary relationships among viruses. Although there is no substitute for sampling infections at earlier points in time, these methods can provide useful information about the genetic makeup of transmitted/founder HIV.

Rare individuals homozygous for a naturally-occurring 32 base pair deletion in the CCR5 gene (CCR5∆32/∆32) are resistant to infection by CCR5-using (“R5”) HIV-1 strains but remain susceptible to less common CXCR4-using (“X4”) strains. The evolutionary dynamics of X4 infections however, remain incompletely understood. We identified two individuals, one CCR5wt/wt and one CCR5∆32/∆32, within the Vancouver Injection Drug Users Study who were infected with a genetically similar X4 HIV-1 strain. While early-stage plasma viral loads were comparable in the two individuals (~4.5–5 log10 HIV-1 RNA copies/ml), CD4 counts in the CCR5wt/wt individual reached a nadir of 3 within 17 months but remained >250 cells/mm3 in the CCR5∆32/∆32 individual. Ancestral phylogenetic reconstructions using longitudinal envelope-V3 deep sequences suggested that both individuals were infected by a single transmitted/founder (T/F) X4 virus that differed at only one V3 site (codon 24). While substantial within-host HIV-1 V3 diversification was observed in plasma and PBMC in both individuals, the CCR5wt/wt individual’s HIV-1 population gradually reverted from 100% X4 to ~60% R5 over ~4 years whereas the CCR5∆32/∆32 individual’s remained consistently X4. Our observations illuminate early dynamics of X4 HIV-1 infections and underscore the influence of CCR5 genotype on HIV-1 V3 evolution.

A phylogeny is a tree-based model of common ancestry that is an indispensable tool for studying biological variation. Phylogenies play a special role in the study of rapidly evolving populations such as viruses, where the proliferation of lineages is constantly being shaped by the mode of virus transmission, by adaptation to immune systems, and by patterns of human migration and contact. These processes may leave an imprint on the shapes of virus phylogenies that can be extracted for comparative study; however, tree shapes are intrinsically difficult to quantify. Here we present a comprehensive study of phylogenies reconstructed from 38 different RNA viruses from 12 taxonomic families that are associated with human pathologies. To accomplish this, we have developed a new procedure for studying phylogenetic tree shapes based on the ‘kernel trick’, a technique that maps complex objects into a statistically convenient space. We show that our kernel method outperforms nine different tree balance statistics at correctly classifying phylogenies that were simulated under different evolutionary scenarios. Using the kernel method, we observe patterns in the distribution of RNA virus phylogenies in this space that reflect modes of transmission and pathogenesis. For example, viruses that can establish persistent chronic infections (such as HIV and hepatitis C virus) form a distinct cluster. Although the visibly ‘star-like’ shape characteristic of trees from these viruses has been well-documented, we show that established methods for quantifying tree shape fail to distinguish these trees from those of other viruses. The kernel approach presented here potentially represents an important new tool for characterizing the evolution and epidemiology of RNA viruses.