Viral quasispecies reconstruction via tensor factorization with successive read removal

Abstract

MotivationAs RNA viruses mutate and adapt to environmental changes, often developing resistance to anti-viral vaccines and drugs, they form an ensemble of viral strains––a viral quasispecies. While high-throughput sequencing (HTS) has enabled in-depth studies of viral quasispecies, sequencing errors and limited read lengths render the problem of reconstructing the strains and estimating their spectrum challenging. Inference of viral quasispecies is difficult due to generally non-uniform frequencies of the strains, and is further exacerbated when the genetic distances between the strains are small.ResultsThis paper presents TenSQR, an algorithm that utilizes tensor factorization framework to analyze HTS data and reconstruct viral quasispecies characterized by highly uneven frequencies of its components. Fundamentally, TenSQR performs clustering with successive data removal to infer strains in a quasispecies in order from the most to the least abundant one; every time a strain is inferred, sequencing reads generated from that strain are removed from the dataset. The proposed successive strain reconstruction and data removal enables discovery of rare strains in a population and facilitates detection of deletions in such strains. Results on simulated datasets demonstrate that TenSQR can reconstruct full-length strains having widely different abundances, generally outperforming state-of-the-art methods at diversities 1–10% and detecting long deletions even in rare strains. A study on a real HIV-1 dataset demonstrates that TenSQR outperforms competing methods in experimental settings as well. Finally, we apply TenSQR to analyze a Zika virus sample and reconstruct the full-length strains it contains.Availability and implementationTenSQR is available at https://github.com/SoYeonA/TenSQR.Supplementary information Supplementary data are available at Bioinformatics online.

A Graph Auto-Encoder for Haplotype Assembly and Viral Quasispecies Reconstruction

“…The sampling is sparse because the reads are much shorter than the haplotypes; moreover, the reads may be erroneous due to sequencing errors. Following (Ahn 2018) , we formalize the sampling operation as…”

“…where M is the one-to-one mapping from the set of reconstructed haplotype to the set of true haplotype (Hashemi 2018), i.e., mapping that determines the best possible match between the two sets of haplotypes. To characterize performance of methods for reconstruction of viral quasispecies with generally a priori unknown number of components, in addition to correct phasing rate we also quantify recall rate, defined as the fraction of perfectly reconstructed components in a population (i.e., recall rate = T P T P +F N ), and predicted proportion, defined as the ratio of the estimated and the true number of components in a genomic mixture (Ahn 2018).…”

“…It is desirable, however, that in the definition of a loss function the distance between numerical values representing any two alleles is identical, no matter which pair of alleles is considered; this ensures the loss function relates to the MEC score -the metric of interest in haplotype assembly problems. Following (Ahn 2018), we define the loss function of the auto-encoder as the squared Frobenius norm of the difference between a one-hot SNP fragment matrix R and the reconstructed matrixR = ZH at the informative positions, i.e., L = 1 2 ||P Ω (R − ZH)|| 2 F , where R ∈ {0, 1} m×4n and H ∈ {0, 1} k×4n are formed by substituting discrete values w ∈ {1, 2, 3, 4} by the set of four dimensional standard basis vectors e (4)…”

“…Determining genetic diversity of a virus is essential for the understanding of its origin and mutation patterns, and the development of effective drug treatments. Reconstructing viral quasispecies (i.e., viral haplotypes, as we refer to them for convenience) is even more challenging than haplotype assembly (Ahn 2018) since the number of constituent strains in a community is typically unknown, and its spectra (i.e., strain frequencies) non-uniform.…”

A binning tool to reconstruct viral haplotypes from assembled contigs

“…Our method is designed to cluster contigs produced by existing assembly tools. There are another group of methods conducting haplotype reconstruction via read clustering [22,23], which groups variant sites obtained by read mapping against reference genomes. These tools don't usually output contigs and thus do not use contig binning.…”