Computational approaches in viral ecology

Abstract

Dynamic virus-host interactions play a critical role in regulating microbial community structure and function. Yet for decades prior to the genomics era, viruses were largely overlooked in microbial ecology research, as only low-throughput culture-based methods of discovering viruses were available. With the advent of metagenomics, culture-independent techniques have provided exciting opportunities to discover and study new viruses. Here, we review recently developed computational methods for identifying viral sequences, exploring viral diversity in environmental samples, and predicting hosts from metagenomic sequence data. Methods to analyze viruses in silico utilize unconventional approaches to tackle challenges unique to viruses, such as vast diversity, mosaic viral genomes, and the lack of universal marker genes. As the field of viral ecology expands exponentially, computational advances have become increasingly important to gain insight into the role viruses in diverse habitats.

Keywords: Bacteriophage-host; Bioinformatics; Microbial ecology; Viral diversity; Viral metagenomics.

Graphical abstract

Fig. 1

Hypothetical viral genomes in squares represented by two types of networks. A) Bipartite network shows which gene families (in circles) are shared by the viral genomes (V1–V4). B) The network in (A) is simplified into a monopartite network between viral genomes with the thickness of the edges representing how many gene families are shared between genomes. This figure was recreated from an article by Iranzo et al. .

Fig. 2

Percentage of correct hosts identified out of 820 assignments by four methods at the genus and species level. This figure was recreated using supplementary data provided by Edwards et al. .