Biology subjects

Poon, A.

Publications and source records attributed to Poon, A..

At least 19 recordsLinked to original sources

bioRxiv2026

Integrated Skin Sensitization Assessment Using Consensus Modelling and Tiered New Approach Methodology: A SaferSkin Case Study

The transition toward animal-free safety assessment of chemicals has accelerated the development of New Approach Methodologies (NAMs) for predicting skin sensitization. However, individual in silico models and experimental NAM assays frequently produce inconsistent or contradictory results, limiting their reliability when used in isolation. To address this challenge, we present a tiered integrated assessment framework implemented through the open source SaferSkin application, which enables systematic comparison and integration of multiple predictive models and experimental data within a transparent weight-of-evidence workflow. In this case study, a diverse set of 21 reference compounds was evaluated using a battery of in silico approaches, including the OECD QSAR Toolbox, VEGA, CASE Ultra and additional machine-learning models implemented within SaferSkin. The platform enables side-by-side comparison of predictions and integration of experimental data through Bayesian network models, allowing probabilistic updating of predictions as new evidence becomes available. Our results demonstrate that reliance on any single predictive model is insufficient for robust hazard identification due to frequent disagreement between models. In contrast, consensus interpretation across multiple modelling approaches combined with targeted experimental evidence substantially improves predictive confidence. The integrated weight-of-evidence framework showed strong concordance with reference classifications and was further supported by independent validation using the Pred-Skin Bayesian model. Importantly, the tiered workflow enables resolution of ambiguous cases. For example, lower-tier predictions for ethyl (2E,4Z)-deca-dienoate were inconsistent across models, whereas targeted third-tier testing using the SENS-IS assay identified the compound as a strong sensitiser (GHS Category 1A). Overall, this study demonstrates how integrated modelling, Bayesian evidence updating and targeted NAM testing can reduce uncertainty in skin sensitization assessment. The SaferSkin framework provides a transparent and reproducible approach for implementing Next Generation Risk Assessment (NGRA) strategies and supports the development of animal-free regulatory toxicology and Safe-and-Sustainable-by-Design chemical innovation. O_FIG O_LINKSMALLFIG WIDTH=200 HEIGHT=115 SRC="FIGDIR/small/711911v1_ufig1.gif" ALT="Figure 1"> View larger version (45K): org.highwire.dtl.DTLVardef@196a52eorg.highwire.dtl.DTLVardef@13690a6org.highwire.dtl.DTLVardef@1528703org.highwire.dtl.DTLVardef@11e05d2_HPS_FORMAT_FIGEXP M_FIG O_FLOATNOGraphical AbstractC_FLOATNO C_FIG

pharmacology and toxicology↗

bioRxiv2025

Evolutionary fingerprinting of protein-coding genes in RNA viruses

RNA viruses evolve rapidly to adapt to changing host environments. Much of this adaptation occurs at the level of protein-coding genes. Thus, some of the most well-characterized examples of rapid adaptation have been found in virus proteins that are exposed on the surface of the viral particle, where they mediate host receptor binding and cell entry. To investigate whether surface-exposed proteins and other proteins encoded by viruses exhibit different patterns of evolution under selection, we analyzed 244 protein-coding genes from 28 species of RNA viruses representing 15 taxonomic families. First, we show that gene-wide rates of non-synonymous (dN) and synonymous (dS) substitutions do not differentiate between categories of proteins. To provide a more detailed comparison between genes, we inferred for each alignment the bivariate posterior distribution over a fixed grid of codon site-specific dN and dS values. This distribution is the genes evolutionary fingerprint. Next, we computed the Wasserstein distance for every pair of fingerprints, which is analogous to amount of work required to reshape one distribution to another. After compensating for differences in genetic variation among viruses and proteins, we found that surface-exposed proteins could not be distinguished from non-exposed proteins in the space induced by the Wasserstein distance matrix. However, surface-exposed proteins from enveloped viruses were significantly clustered apart from their counterparts in non-enveloped viruses. In contrast, there was no significant separation between these categories of viruses for proteins with polymerase activity. We show that this pattern is more consistent with relaxed purifying selection than adaptive evolution in proteins associated with viral envelopes.

evolutionary biology↗

bioRxiv2025

Phylogenetic discordance cannot identify genomic reassortment events: a case study with the avian influenza virus H5Nx

Recombination plays an important role in the evolution of RNA viruses, as it allows the exchange of genetic material between viral lineages. Reassortment, a form of recombination specific to segmented genomes, involves the exchange of entire segments and has contributed significantly to the adaptation and spread of influenza viruses through novel genomic combinations, i.e., antigenic shifts. It is usually identified by phylogenetic discordance: differences in the topologies of trees reconstructed from different genomic segments. However, phylogenetic discordance can also result from error in reconstructing trees. To characterize the impact of reconstruction error, we curated a database of n = 11, 765 complete genomes of avian H5Nx influenza A viruses from avian hosts. We found evidence of widespread reassortment as measured by inferred subtree-prune-regraft (SPR) events, consistent with previous studies. Next, we ran replicate simulations of sequence evolution along the reference tree for the segment encoding hemagglutinin (HA), adjusting simulations for the lengths and clock rates of the other segments. These simulations provided a baseline for the expected amount of phylogenetic discordance in the absence of any reassortment. When sampling HA sequences at random from the database to build reference trees, we observed that simulating other segments without reassortment still yielded about 32% as many SPRs as the real segment data on average. The average proportion of SPRs without reassortment was greatly reduced (4%) if we selected an equivalent number of HA sequences retaining the most genetic diversity, which was consistent with the accuracy of phylogenetic reconstruction being the limiting factor. This implies that measuring reassortment by SPRs may have a high false positive rate, and that previous evidence of extensive reassortment in influenza viruses should be interpreted with caution. In addition, we observed that the SPRs reconstructed on simulated trees had significantly shorter distances between the prune and regraft locations than real trees. These results suggest that down-sampling sequences to maximize evolutionary divergence and filtering out the shortest SPRs may be effective measures against false positives.

bioinformatics↗

bioRxiv2024

Skin Sensitisation Case Study: Comparison of Defined Approaches including OECD 497 Guidance

Characterising known and new chemical compounds for skin sensitisation provides a basis for the development of safer products where ingredients are exposed to skin. By including new approaches, such as tiered testing strategies and integrated data analysis, it is possible to develop next generation products adhering to emerging regulations, scientific evidence and animal welfare principles. To ensure data integrity during such assessment the OECD provides characterisation guidelines and Defined Approaches (DAs) to uniform work-flows. In this study we developed and applied the integrated characterisation tool called < > to compare the results of different DAs for eight compounds and included results obtained from current OECD guidance and emerging methods. We tested two compounds with unclear or indeterminate results with the SENS-IS assay to explore the value of the experiment in strengthening the weight of evidence and arriving at a clearer conclusion.

pharmacology and toxicology↗

bioRxiv2023

Within-host rates of insertion and deletion in the HIV-1 surface envelope glycoprotein

Under selection by neutralizing antibodies, the HIV-1 envelope glycoprotein gp120 undergoes rapid evolution within hosts, particularly in regions encoding the five variable loops (V1-V5). Indel polymorphisms are abundant in these loops, where they can facilitate immune escape by modifying the length, composition and glycosylation profile of these structures. Here, we present a comparative analysis of within-host indel rates and characteristics within the variable regions of gp120. We analyzed a total of 3,437 HIV-1 gp120 sequences sampled longitudinally from 29 different individuals using coalescent models in BEAST. Next, we used Historian to reconstruct ancestral sequences from the resulting tree samples, and fit a Poisson generalized linear model to the distribution of indel events to estimate their rates in the five variable loops. Overall, the mean insertion and deletion rates were 1.6 x 10-3 and 2.5 x 10-3/ nt / year, respectively, with significant variation among loops. Insertions and deletions also followed similar length distributions, except for significantly longer indels in V1 and V4 and shorter indels in V5. Insertions in V1, V2, and V4 tended to create new N-linked glycosylation sites significantly more often than expected by chance, which is consistent with positive selection to alter glycosylation patterns.

evolutionary biology↗

bioRxiv2023

Prospects for a sequence-based taxonomy of influenza A virus subtypes

The hemagglutinin (HA) and neuraminidase (NA) proteins are the primary antigenic targets of influenza A virus (IAV) infections. IAV infections are generally classified into subtypes of HA and NA proteins, e.g., H3N2. Most of the known subtypes were originally defined by a lack of antibody cross-reactivity. However, genetic sequencing has played an increasingly important role in characterizing the evolving diversity of IAV. Novel subtypes have recently been described solely by their genetic sequences, and IAV infections are routinely subtyped by molecular assays, e.g., real-time PCR, or the comparison of sequences to references. In this study, I carry out a phylogenetic analysis of all available IAV protein sequences in the Genbank database (over 1.1 million records) to determine whether the serologically-defined subtypes can be reproduced with sequence-based criteria. I show that a robust genetic taxonomy of HA and NA subtypes can be obtained with a simple clustering method, namely by progressively partitioning the phylogeny on its longest internal branches. However, this taxonomy also requires some amendments to the current nomenclature. For example, two IAV isolates from bats previously characterized as a divergent lineage of H9N2 should be separated into their own subtype. With the exception of these small and highly divergent lineages, the phylogenies relating each of the other six genomic segments do not support partitions into major subtypes.

bioinformatics↗

bioRxiv2022

Golgi Fragmentation - One of the Earliest Organelle Phenotypes in Alzheimer's Disease Neurons

Alzheimers disease (AD) is the most common cause of dementia, with no current cure. Consequently, alternative approaches focusing on early pathological events in specific neuronal populations, besides targeting the well-studied Amyloid beta (A{beta}) accumulations and Tau tangles, are needed. In this study, we have investigated disease phenotypes specific to glutamatergic forebrain neurons and mapped the timeline of their occurrence, by implementing familial and sporadic human induced pluripotent stem cell models as well as the 5xFAD mouse model. We recapitulated characteristic late AD disease phenotypes, such as increased A{beta} secretion and Tau hyperphosphorylation, as well as previously well documented mitochondrial and synaptic deficits. Intriguingly, we identified Golgi fragmentation as one of the earliest AD phenotypes, indicating potential impairments in protein processing and post-translational modifications. Computational analysis of RNA sequencing data revealed differentially expressed genes involved in glycosylation and glycan patterns, whilst total glycan profiling revealed minor glycosylation differences. This indicates general robustness of glycosylation besides the observed fragmented morphology. Importantly, we identified that genetic variants in Sortilin-related receptor 1 (SORL1) associated with AD could aggravate the Golgi fragmentation and subsequent glycosylation changes. In summary, we identified Golgi fragmentation as one of the earliest disease phenotypes in AD neurons in various in vivo and in vitro complementary disease models, which can be exacerbated via additional risk variants in SORL1. O_FIG O_LINKSMALLFIG WIDTH=200 HEIGHT=140 SRC="FIGDIR/small/519571v1_ufig1.gif" ALT="Figure 1"> View larger version (39K): org.highwire.dtl.DTLVardef@98a8eeorg.highwire.dtl.DTLVardef@7aa7b7org.highwire.dtl.DTLVardef@991e33org.highwire.dtl.DTLVardef@8dabb4_HPS_FORMAT_FIGEXP M_FIG C_FIG

neuroscience↗

bioRxiv2022

HexSE: Simulating evolution in overlapping reading frames

MotivationGene overlap occurs when two or more genes are encoded by the same nucleotides. This phenomenon is found in all taxonomic domains, but is particularly common in viruses, where it may provide a mechanism to increase the information content of compact genomes. The presence of overlapping reading frames (OvRFs) can skew estimates of selection based on the rates of non-synonymous and synonymous substitutions, since a substitution that is synonymous in one reading frame may be non-synonymous in another, and vice versa. ResultsTo understand the impact of OvRFs on molecular evolution, we implemented a versatile simulation model of nucleotide sequence evolution along a phylogeny with any distribution of open reading frames in linear or circular genomes. We use a custom data structure to track the substitution rates at every nucleotide site, which is determined by the stationary nucleotide frequencies, transition bias, and the distribution of selection biases (dN/dS) in the respective reading frames. Availability and implementationOur simulation model is implemented in the Python scripting language. All source code is released under the GNU General Public License (GPL) version 3, and is available at https://github.com/PoonLab/HexSE.

bioinformatics↗

bioRxiv2022

From components to communities: bringing network science to clustering for genomic epidemiology

Defining clusters of epidemiologically-related infections is a common problem in the surveillance of infectious disease. A popular method for generating clusters is pairwise distance clustering, which assigns pairs of sequences to the same cluster if their genetic distance falls below some threshold. The result is often represented as a network or graph of infections. A connected component is a set of interconnected nodes in a graph that are not connected to any other node. The current approach to pairwise clustering is to map clusters to the connected components of the graph. However, the distance thresholds typically used for viruses like HIV-1 tend to yield components that exclude large numbers of infections as unconnected nodes. This is problematic for public health applications of clustering, such as tracking the growth of clusters over time. We propose that this problem can be addressed with community detection, a class of clustering methods being developed in the field of network science. A community is a set of nodes that are more densely inter-connected relative to the number of connections to external nodes. Thus, a connected component may be partitioned into two or more communities. Here we describe community detection methods in the context of genetic clustering for epidemiology, demonstrate how a popular method (Markov clustering) enables us to resolve variation in transmission rates within a giant connected component of HIV-1 sequences, and identify current challenges and directions for further work.

bioinformatics↗

bioRxiv2022

Optimized phylogenetic clustering of HIV-1 sequence data for public health applications

Clusters of genetically similar infections suggest rapid transmission and may indicate priorities for public health action or reveal underlying epidemiological processes. However, clusters often require user-defined thresholds and are sensitive to non-epidemiological factors, such as non-random sampling. Consequently the ideal threshold for public health applications varies substantially across settings. Here, we show a method which selects optimal thresholds for phylogenetic (subset tree) clustering based on population. We evaluated this method on HIV-1 pol datasets (n = 14,221 sequences) from four sites in USA (Tennessee, Seattle), Canada (Northern Alberta) and China (Beijing). Clusters were defined by tips descending from an ancestral node (with a minimum bootstrap support of 95%) through a series of branches, each with a length below a given threshold. Next, we used pplacer to graft new cases to the fixed tree by maximum likelihood. We evaluated the effect of varying branch-length thresholds on cluster growth as a count outcome by fitting two Poisson regression models: a null model that predicts growth from cluster size, and an alternative model that includes mean collection date as an additional covariate. The alternative model was favoured by AIC across most thresholds, with optimal (greatest difference in AIC) thresholds ranging 0.007-0.013 across sites. The range of optimal thresholds was more variable when re-sampling 80% of the data by location (IQR 0.008 - 0.016, n = 100 replicates). Our results use prospective phylogenetic cluster growth and suggest that there is more variation in effective thresholds for public health than those typically used in clustering studies.

bioinformatics↗

bioRxiv2021

CoVizu: Rapid analysis and visualization of the global diversity of SARS-CoV-2 genomes

Phylogenetics has played a pivotal role in the genomic epidemiology of SARS-CoV-2, such as tracking the emergence and global spread of variants, and scientific communication. However, the rapid accumulation of genomic data from around the world -- with over two million genomes currently available in the GISAID database -- is testing the limits of standard phylogenetic methods. Here, we describe a new approach to rapidly analyze and visualize large numbers of SARS-CoV-2 genomes. Using Python, genomes are filtered for problematic sites, incomplete coverage, and excessive divergence from a strict molecular clock. All differences from the reference genome, including indels, are extracted using minimap2, and compactly stored as a set of features for each genome. For each Pango lineage (https://cov-lineages.org), we collapse genomes with identical features into variants, generate 100 bootstrap samples of the feature set union to generate weights, and compute the symmetric differences between the weighted feature sets for every pair of variants. The resulting distance matrices are used to generate neigihbor-joining trees in RapidNJ and converted into a majority-rule consensus tree for the lineage. Branches with support values below 50% or mean lengths below 0.5 differences are collapsed, and tip labels on affected branches are mapped to internal nodes as directly-sampled ancestral variants. Currently, we process about million genomes in approximately nine hours on 34 cores. The resulting trees are visualized using the JavaScript framework D3.js as beadplots, in which variants are represented by horizontal line segments, annotated with beads representing samples by collection date. Variants are linked by vertical edges to represent branches in the consensus tree. These visualizations are published at https://filogeneti.ca/CoVizu. All source code was released under an MIT license at https://github.com/PoonLab/covizu.

bioinformatics↗

bioRxiv2021

Clustering and visualizing the distribution of overlapping reading frames in virus genomes

Gene overlap occurs when two or more genes are encoded by the same nucleotides. This phenomenon is found in all taxonomic domains, but is particularly common in viruses, where it may increase the information content of compact genomes or influence the creation of new genes. Here we report a global comparative study of overlapping reading frames (OvRFs) of 12,609 virus reference genomes in the NCBI database. We retrieved metadata associated with all annotated reading frames in each genome record to calculate the number, length, and frameshift of OvRFs. Our results show that while the number of OvRFs increases with genome length, they tend to be shorter in longer genomes. The majority of overlaps involve +2 frameshifts, predominantly found in ds-DNA viruses. However, the longest overlaps involve no shift in reading frame (+0), increasing the selective burden of the same nucleotide positions within codons, instead of exposing additional sites to purifying selection. Next, we develop a new graph-based representation of the distribution of OvRFs among the reading frames of genomes in a given virus family. In the absence of an unambiguous partition of reading frames by homology at this taxonomic level, we used an alignment-free k-mer based approach to cluster protein coding sequences by similarity. We connect these clusters with two types of directed edges to indicate (1) that constituent reading frames are adjacent in one or more genomes, and (2) that the reading frames overlap. These adjacency graphs not only provide a natural visualization scheme, but also a novel statistical framework for analyzing the effects of gene- and genome-level attributes on the frequencies of overlaps.

bioinformatics↗

bioRxiv2019

Public health in genetic spaces: a statistical framework to optimize cluster-based outbreak detection

Genetic clustering is a popular method for characterizing variation in transmission rates for rapidly-evolving viruses, and could potentially be used to detect outbreaks in near real time. However, the statistical properties of clustering are poorly understood in this context, and there are no objective guidelines for setting clustering criteria. Here we develop a new statistical framework to optimize a genetic clustering method based on the ability to forecast new cases. We analyzed the pairwise Tamura-Nei (TN93) genetic distances for anonymized HIV-1 subtype B pol sequences from Seattle (n = 1, 653) and Middle Tennessee, USA (n = 2, 779), and northern Alberta, Canada (n = 809). Under varying TN93 thresholds, we fit two models to the distributions of new cases relative to clusters of known cases: (1) a null model that assumes cluster growth is strictly proportional to cluster size, i.e., no variation in transmission rates among individuals; and (2) a weighted model that incorporates individual-level covariates, such as recency of diagnosis. The optimal threshold maximizes the difference in information loss between models, where covariates are used most effectively. Optimal TN93 thresholds varied substantially between data sets, e.g., 0.0104 in Alberta and 0.016 in Seattle and Tennessee, such that the optimum for one population will potentially mis-direct prevention efforts in another. The range of thresholds where the weighted model conferred greater predictive accuracy tended to be narrow ({+/-}0.005 units), but the optimal threshold for a given population also tended to be stable over time. We also extended our method to demonstrate that variation in recency of HIV diagnosis among clusters was significantly more predictive of new cases than sample collection dates ({Delta}AIC> 50). These results demonstrate that one cannot rely on historical precedence or convention to configure genetic clustering methods for public health applications. Our framework not only provides an objective procedure to optimize a clustering method, but can also be used for variable selection in forecasting new cases.

epidemiology↗

bioRxiv2019

Network science inspires novel tree shape statistics

1The shape of phylogenetic trees can be used to gain evolutionary insights. A trees shape specifies the connectivity of a tree, while its branch lengths reflect either the time or genetic distance between branching events; well-known measures of tree shape include the Colless and Sackin imbalance, which describe the asymmetry of a tree. In other contexts, network science has become an important paradigm for describing structural features of networks and using them to understand complex systems, ranging from protein interactions to social systems. Network science is thus a potential source of many novel ways to characterize tree shape, as trees are also networks. Here, we tailor tools from network science, including diameter, average path length, and betweenness, closeness, and eigenvector centrality, to summarize phylogenetic tree shapes. We thereby propose tree shape summaries that are complementary to both asymmetry and the frequencies of small configurations. These new statistics can be computed in linear time and scale well to describe the shapes of large trees. We apply these statistics, alongside some conventional tree statistics, to phylogenetic trees from three very different viruses (HIV, dengue fever and measles), from the same virus in different epidemiological scenarios (influenza A and HIV) and from simulation models known to produce trees with different shapes. Using mutual information and supervised learning algorithms, we find that the statistics adapted from network science perform as well as or better than conventional statistics. We describe their distributions and prove some basic results about their extreme values in a tree. We conclude that network science-based tree shape summaries are a promising addition to the toolkit of tree shape features. All our shape summaries, as well as functions to select the most discriminating ones for two sets of trees, are freely available as an R package at http://github.com/Leonardini/treeCentrality.

bioinformatics↗

bioRxiv2018

Phylogenetic measures of indel rate variation among the HIV-1 group M subtypes

The transmission and pathogenesis of human immunodeficiency virus type 1 (HIV-1) is disproportionately influenced by evolution in the five variable regions of the virus surface envelope glycoprotein (gp120). Insertions and deletions (indels) are a significant source of evolutionary change in these regions. However, the influx of indels relative to nucleotide substitutions has not yet been quantified through a comparative analysis of HIV-1 sequence data. Here we develop and report results from a phylogenetic method to estimate indel rates for the gp120 variable regions across five major subtypes and two circulating recombinant forms (CRFs) of HIV-1 group M. We processed over 26,000 published HIV-1 gp120 sequences, from which we extracted 6,605 sequences for phylogenetic analysis. In brief, our method employs maximum likelihood to reconstruct phy-logenies scaled in time and fits a Poisson model to the observed distribution of indels between closely related pairs of sequences in the tree (cherries). The rate estimates ranged from 3.0 x 10-5 to 1.5 x 10-3 indels/nt/year and varied significantly among variable regions and subtypes. Indel rates were significantly lower in the region encoding variable loop V3, and also lower for HIV-1 subtype B relative to other subtypes. We also found that variable loops V1, V2 and V4 tended to accumulate significantly longer indels. Further, we observed that the nucleotide composition of indel sequences was significantly distinct from that of the flanking sequence in HIV-1 gp120. Indels affected potential N-linked glycosylation sites substantially more often in V1 and V2 than expected by chance, which is consistent with positive selection on glycosylation patterns within these regions of gp120. These results represent the first comprehensive measures of indel rates in HIV-1 gp120 across multiple subtypes and CRFs, and identifies novel and unexpected patterns for further research in the molecular evolution of HIV-1.

microbiology↗

bioRxiv2018

An open-source k-mer based machine learning tool for fast and accurate subtyping of HIV-1 genomes

For many disease-causing virus species, global diversity is clustered into a taxonomy of subtypes with clinical significance. In particular, the classification of infections among the subtypes of human immunodeficiency virus type 1 (HIV-1) is a routine component of clinical management, and there are now many classification algorithms available for this purpose. Although several of these algorithms are similar in accuracy and speed, the majority are proprietary and require laboratories to transmit HIV-1 sequence data over the network to remote servers. This potentially exposes sensitive patient data to unauthorized access, and makes it impossible to determine how classifications are made and to maintain the data provenance of clinical bioinformatic workflows. We propose an open-source supervised and alignment-free subtyping method (KO_SCPCAPAMERISC_SCPCAP) that operates on k-mer frequencies in HIV-1 sequences. We performed a detailed study of the accuracy and performance of subtype classification in comparison to four state-of-the-art programs. Based on our testing data set of manually curated real-world HIV-1 sequences (n = 2, 784), Kameris obtained an overall accuracy of 97%, which matches or exceeds all other tested software, with a processing rate of over 1,500 sequences per second. Furthermore, our fully standalone general-purpose software provides key advantages in terms of data security and privacy, transparency and reproducibility. Finally, we show that our method is readily adaptable to subtype classification of other viruses including dengue, influenza A, and hepatitis B and C virus.

bioinformatics↗

bioRxiv2018

Food web networks shift across a precipitation gradient due to changes in community composition and species interactions.

Ecological networks change across spatial and environmental gradients due to (i) changes in species composition or (ii) changes in the frequency or strength of interactions. Here we use the communities of aquatic invertebrates inhabiting clusters of bromeliad phytotelms along the Brazilian coast as a model system for examining turnover in the properties of ecological networks. We first document the variation in the species pools of sites across a geographical climate gradient. Using the same sites, we also explored the geographic variation in species interaction strength using a newly developed Markov network approach. We found that community composition differed along a gradient of water volume within bromeliads due to the turnover of some species. From the Markov network analysis, we found that the top-down effects of certain predators differed geographically, which could also be explained by geographic differences in bromeliad water volumes. Overall, this study illustrates how a network can change across an environmental gradient through both changes in both species and their interactions.

ecology↗

bioRxiv2017

A model-based clustering method to detect infectious disease transmission outbreaks from sequence variation

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 HIV-1 sequences from a study cohort of men who have sex with men in Seattle, USA, we found that the MMPP method categorized about half (46%) as many individuals to clusters compared to the other methods, and that the MMPP clusters were more consistent with transmission outbreaks. 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 resources.

epidemiology↗