CorGAT: a tool for the functional annotation of SARS-CoV-2 genomes

Abstract

Summary

While over 200 000 genomic sequences are currently available through dedicated repositories, ad hoc methods for the functional annotation of SARS-CoV-2 genomes do not harness all currently available resources for the annotation of functionally relevant genomic sites. Here, we present CorGAT, a novel tool for the functional annotation of SARS-CoV-2 genomic variants. By comparisons with other state of the art methods we demonstrate that, by providing a more comprehensive and rich annotation, our method can facilitate the identification of evolutionary patterns in the genome of SARS-CoV-2.

Availabilityand implementation

Galaxy

 

http://corgat.cloud.ba.infn.it/galaxy; software: https://github.com/matteo14c/CorGAT/tree/Revision_V1; docker: https://hub.docker.com/r/laniakeacloud/galaxy_corgat.

Supplementary information

Supplementary data are available at Bioinformatics online.

1 Introduction

The recent outbreak of COVID-19 has underlined the importance of rapid and effective sharing of molecular data for combating the spread of human pathogens and tracing possible routes of infection. At present, more than 200 000 genomic SARS-CoV-2 sequences have been deposited in dedicated repositories (Shu and McCauley, 2017) along with associated metadata. Harnessing this wealth of information to identify functionally relevant genomic changes and/or recognizing the emergence of novel viral strains is of pivotal importance in the fight against COVID-19. Currently tools for functional annotation of genomic sequences have not been specifically devised for the analysis of SARS-CoV-2, e.g. do not take into account the unusual mechanisms of transcription and post-translational processing of coronaviruses gene products (Sawicki et al., 2007). Additionally, while a wealth of resources and datasets for the fine-grained annotation of functional genomic elements are currently available, including: detailed studies of transcriptional mechanisms (Kim et al., 2020), conserved regulatory sequences (Sawicki et al., 2007), sites under evolutionary selection (http://hyphy.org/covid/), predicted epitopes (Kiyotani et al., 2020) and non-coding secondary structure elements, are available, these are not normally incorporated in the functional annotation of SARS-CoV-2 genomic variants. To overcome these limitations, we propose a novel, highly effective and user friendly approach for the functional annotation of SARS-CoV-2 genomes: CorGAT - the Coronavirus Genome Analysis Tool. By integrating a curated selection of datasets and resources, CorGAT provides a richer and more detailed annotation of SARS-CoV-2 variants when compared with other state of the art methods. To illustrate its advantages, we apply CorGAT to the complete collection of 57 558 currently available SARS-CoV-2 genomic sequences, and derive relevant insights concerning the evolution of this novel pathogen.

2 Implementation

CorGAT has been made available as a collection of Perl script and annotation files at https://github.com/matteo14c/CorGAT/tree/Revision_V1. A user friendly version of the software is available in the form of a standalone Galaxy (Afgan et al., 2018) implementation, based on the Laniakea@ReCaS Galaxy on-demand service (Tangaro et al., 2020) at http://corgat.cloud.ba.infn.it/galaxy. A Docker container image can be obtained from https://hub.docker.com/r/laniakeacloud/galaxy_corgat. A complete account of the resources used for the annotation of the SARS-CoV-2 genomes, and of their integration in CorGAT is presented in the Supplementary Materials. A detailed user manual is available at https://corgat.readthedocs.io/. Functional annotation files incorporated in CorGAT are updated on a monthly basis, to cope with the constant increase in publicly available data and genomic sequences of SARS-CoV-2. CorGAT has a modular architecture (see Supplementary Materials), allowing the rapid inclusion of novel or even custom types of annotations, simply by editing plain text files.

3 Results

To demonstrate the application of CorGAT, we compared the functional annotation of the complete collection of 20 045 genetic variants derived from 57 558 genomic sequences of SARS-CoV-2 (see Supplementary Materials) by CorGAT, with the annotations by SNPeff (Cingolani et al., 2012) and by the Variant Annotation Integrator (Hinrichs et al., 2016). Simple statistics concerning the number and types of variants are reported in Supplementary Table S1. As outlined in Supplementary Table S2, all the tools herein considered provided highly consistent annotations of functional effects of variants associated with protein coding genes, thus confirming that CorGAT attains the same level of sensitivity as the other methods. However, as illustrated in Supplementary Table S3, CorGAT provides additional layers of annotation that are not provided by other methods, for a total of 14753 single distinct annotations. These include 33 variants associated with regulatory elements (transcription regulatory sequences, TRS), 69 variants associated to consensus cleavage sites (Kiemer et al., 2004) in the ORF1a and ORF1ab polyproteins, 1164 variants associated with sites under selection according to Hyphy (Kosakovsky Pond et al., 2020) and 161 variants in conserved secondary structure elements (Supplementary Table S3). According to our analyses, a highly significant reduction of missense substitutions is observed at sites predicted to be under negative selection (Fisher P-value < 2.2e-16), compared to the background of all the substitutions in protein coding genes. Nevertheless, 229 missense substitutions alter highly conserved amino acid residues that are predicted to be under negative selection. Furthermore, analysis of genetic variants associated with functional non-coding elements in the genome of SARS-CoV-2 highlight some potentially interesting patterns. While the 5′ and 3′ UTRs are the most variable regions of the genome, TRS and secondary structure elements in general are considerably less variable, and show levels of conservation comparable to protein coding genes (Supplementary Table S4). This is well exemplified by the TRS-L element, which is the single most conserved region in the 5′ UTR (Supplementary Fig.  S1). Strikingly, the s2m element in the 3′ UTR (Tengs et al., 2013) exhibits more of variability and recurrent indels than other annotated functional elements (Supplementary Tables S4 and S5). Interestingly, our functional annotation (see Supplementary Materials), indicates that several observed substitutions might result in substantial changes to s2m structure consistent with change or loss of s2m function in SARS-CoV-2 (Chiara et al., 2020). We conclude, that CorGAT constitutes a useful addition to the collection of tools for the functional characterization of SARS-CoV-2 genomes.