The International Virus Bioinformatics Meeting 2022

Abstract

The International Virus Bioinformatics Meeting 2022 took place online, on 23-25 March 2022, and has attracted about 380 participants from all over the world. The goal of the meeting was to provide a meaningful and interactive scientific environment to promote discussion and collaboration and to inspire and suggest new research directions and questions. The participants created a highly interactive scientific environment even without physical face-to-face interactions. This meeting is a focal point to gain an insight into the state-of-the-art of the virus bioinformatics research landscape and to interact with researchers in the forefront as well as aspiring young scientists. The meeting featured eight invited and 18 contributed talks in eight sessions on three days, as well as 52 posters, which were presented during three virtual poster sessions. The main topics were: SARS-CoV-2, viral emergence and surveillance, virus-host interactions, viral sequence analysis, virus identification and annotation, phages, and viral diversity. This report summarizes the main research findings and highlights presented at the meeting.

Keywords: SARS-CoV-2; bioinformatics; phages; tools; viral diversity; viral emergence and surveillance; viral sequence analysis; virus identification and annotations; virus–host interactions.

Figure 1

Main steps of the VOCAL pipeline. To generate mutation profiles for the spike gene for each input sequence, VOCAL can preprocess raw genome sequences or can directly receive this information from the covSonar.

Figure 2

The power of SARS-CoV-2 genotyping and SNP-based clustering for contextual outbreak assessment. (A) We expect the viral genomes to accumulate mutations, as they are transmitted from one individual to the next. An efficient method to identify chains of genetically similar sequences would therefore be useful to identify putative outbreaks. (B) Diagram illustrating the steps of the clustering algorithm using a maximum SNP distance of 1. The mutation profiles are obtained by a reference-based alignment with Nextclade [17] or covSonar (https://github.com/rki-mf1/covsonar (accessed on 3 May 2022)). The pairwise distances between different sequences are derived from the constructed distance matrix of genomic profiles. Two sequences are part of the same transmission cluster if the pairwise distance between them is below the user-specified threshold, max-dist. The final transmission clusters can be further analyzed with phylogenetic software.

Figure 3

Maximum likelihood phylogenetic tree. Maximum likelihood phylogenetic tree indicating the different clades corresponding to the fifteen major lineages and showing where the query sequence (DVS-321) clusters in the tree. Pairwise distance measure for the different lineages and query samples indicate a genetic diversity, which indicates a maximum diversity of 5% at the nucleotide level.

Figure 4

Virus–host factor interactions impact the gene expression of the host cell. Single-stranded RNA (ssRNA) viruses enter the cell and release their RNA genomes into the cytoplasm of the host cell. Viral RNAs can contain binding sequences for host RNA binding proteins (RBPs). The binding of host RBPs to the viral RNA can have proviral or antiviral effects. The sequestration of RBPs by cytoplasmic viral RNA was reported to cause changes in host cell RNA splicing, polyadenylation, and stability [25].

Figure 5

Origins and implications of the quasispecies concept. Images of the work in Charles Weissmann’s laboratory in Zürich in the 1970s. (Left): a slide drawn by Weissmann outlining a reversion experiment by site-directed mutagenesis [33]; it is interesting that the N4-hydroxy-CTP used as mutagenic nucleotide is the active component of molnupiravir, presently used as lethal mutagen for SARS-CoV-2. (Right): a page of the notebook of Domingo with the experimental data and mathematical predictions of competition between the wild type Q$\beta$ phage and the infectious extracistronic mutant (top of page), and reversion of the mutant upon multiplication in E. coli (bottom of page) [31,32]; explained also in reference [35].

Figure 6

Overview of the viral sequence prediction pipeline used in VirSorter2 [40]. The “hmmDB” includes viral protein HMMs from two of the largest databases, VPF and Efam [38,42]. Distinct classifiers (random forest) are built for each of five major viral groups to improve accuracy on diverse viruses. Adapted with permission from [40] (https://creativecommons.org/licenses/by/4.0/). Copyright 2021, Guo et al.

Figure 7

Dual identification of novel phage receptor-binding proteins. Graphical abstract of the collected phage genome data, the developed RBP detection tools and the benchmark against the recently developed PhANNs tool [50].

Figure 8

Models predicting capsid architecture. (A) The genome to capsid (G2C) model relies on the conserved properties of tailed bacteriophages. The model uses the genome length to predict the capsid architecture (diameter and icosahedral T-number). (B) The major capsid protein to capsid (MCP2C) relies on the G2C model to build a library of putative capsid architectures and MCPs from isolated tailed phage genomes. It predicts the capsid architecture of tailed phages directly from the major capsid protein sequence. The current G2C and MCP2C python versions are accessible at https://github.com/luquelab/Lee_etal_CSBJ_2022/tree/main/3_executables (accessed on 1 May 2022).

Figure 9

(A) Principal coordinates analysis of inter-individual differences of the gut virome (genus level Bray-Curtis dissimilarity) in the IBD cohort (circles colored by viral community-type, $n=366$). (B) Metadata variables significantly correlating to virome compositional variation in both the IBD cohort (left) and post-intervention samples (right) (dbRDA, genus-level Bray-Curtis dissimilarity), as determined by a multivariate linear regression model. (C–E) Modeling the association between the metadata drivers of post-intervention samples and the prevalence of viral community-type CrM (logistic regression, $n=166$, only significant associations shown). (C) Relative risk ratio of prevalence of viral community-type CrM with the significant driver (endoscopic remission) of virome variation. (D) Representation of viral community-type prevalence in post-intervention samples ($n=166$) stratified according to the endoscopic outcome (non-remission, remission, unknown). (E) Representation of viral community-type prevalence in post-intervention samples per IBD subtype, CD (left, $n=115$) and UC (right, $n=51$) stratified according to the endoscopic outcome (non-remission, remission, or unknown). * $p<0.05$ (adjustment for multiple testing was performed using the Benjamini-Hochberg methods).