Genotyping SARS-CoV-2 Through an Interactive Web Application

The ongoing COVID-19 pandemic is the greatest health-care challenge of this generation. Early viral genome sequencing studies of small cohorts have indicated the possibility of distinct severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) genotypes.1Tang X Wu C Li X et al.On the origin and continuing evolution of SARS-CoV-2.Natl Sci Rev. 2020; (published online March 3.)DOI:10.1093/nsr/nwaa036Crossref PubMed Scopus (1059) Google Scholar If these subtypes result in an altered virus tropism or pathogenesis in infected hosts, this could have immediate implications for vaccine design, drug development, and efforts to control the pandemic. Therefore, the genomic surveillance and characterisation of circulating viral strains is a high priority for research and development. To facilitate the epidemiological tracking of SARS-CoV-2, researchers worldwide have created various web-portals and tools, such as the Johns Hopkins University COVID-19 dashboard.2Dong E Du H Gardner L An interactive web-based dashboard to track COVID-19 in real time.Lancet Infect Dis. 2020; 20: 533-534Summary Full Text Full Text PDF PubMed Scopus (6279) Google Scholar An unprecedented effort to make COVID-19-related data accessible in near real-time has resulted in more than 25 000 publicly available genome sequences of SARS-CoV-2 on Global Initiative on Sharing All Influenza Data (GISAID).3Elbe S Buckland-Merrett G Data, disease and diplomacy: GISAID's innovative contribution to global health.Glob Chall. 2017; 1: 33-46Crossref PubMed Scopus (1055) Google Scholar Although platforms to survey epidemiological data are prevalent, tools that summarise publicly available viral genome data are scarce and those that are available do not offer users the ability to analyse in-house sequencing data. To address this gap, we have developed an accessible application, the COVID-19 Genotyping Tool (CGT). A video demonstration of CGT is available in appendix 1. CGT uses publicly deposited SARS-CoV-2 consensus genome sequences from the GISAID EpiCoV™ database,3Elbe S Buckland-Merrett G Data, disease and diplomacy: GISAID's innovative contribution to global health.Glob Chall. 2017; 1: 33-46Crossref PubMed Scopus (1055) Google Scholar and summarises relevant information through salient visualisations, including Uniform Manifold Approximation and Projection (UMAP), Minimum Spanning Trees (MST) of sequence networks, and allele frequencies of annotated high-prevalence non-synonymous Single-Nucleotide Polymorphisms (SNPs) within structural protein-coding genomic regions (envelope, membrane, nucleocapsid, and spike proteins; figure). New sequencing data from GISAID are added to CGT once a week. Currently, three metadata types can be overlaid on the visualisations: the region of sample collection, the country of sample collection, and the sample collection date with respect to the start of the pandemic (heuristically defined as Dec 1, 2019; figure; appendix 2). Users can upload post-assembly consensus FASTA sequences of SARS-CoV-2 for interactive analysis. The 3' and 5' untranslated regions of genomes are trimmed because of low sequence identity, caused by difficulties in their amplification and sequencing. After sequence alignment, the DNA distance is calculated using the Kimura-80 model of nucleotide substitution.4Kimura M A simple method for estimating evolutionary rates of base substitutions through comparative studies of nucleotide sequences.J Mol Evol. 1980; 16: 111-120Crossref PubMed Scopus (24466) Google Scholar After the calculation of the distance and annotation of SNPs, visualisations are reactively reprocessed with user-uploaded data. We use UMAP and network analysis instead of phylogenetics because of a faster computation time and ease of interpretability for large datasets. UMAP reduces the high-dimensional representation of DNA sequences to a two-dimensional embedding, indicating a sequence similarity based on the distance between points;5McInnes L Healy J Melville J UMAP: Uniform Manifold Approximation and Projection for dimension reduction.arXiv. 2018; (published online Feb 9.) (preprint).https://arxiv.org/abs/1802.03426Google Scholar that is, SARS-CoV-2 genomes. The MST of the SARS-CoV-2 genome network is a metric used to define a network subset such that all the nodes are connected while minimising distance.6Mamun A, Rajasekaran S. An efficient Minimum Spanning Tree algorithm. 2016 IEEE Symposium on Computers and Communication; Messina, Italy; June 27–30, 2016 (abstr 1047–52).Google Scholar MSTs have been used before in outbreak analysis to identify the most probable transmission events between hosts,7Spada E Sagliocca L Sourdis J et al.Use of the Minimum Spanning Tree model for molecular epidemiological investigation of a nosocomial outbreak of hepatitis C virus infection.J Clin Microbiol. 2004; 42: 4230-4236Crossref PubMed Scopus (39) Google Scholar and can therefore offer epidemiological insight into SARS-CoV-2 transmission. Lastly, heterogeneity within structural proteins is particularly notable in terms of the host immune response, and has direct implications for vaccine development. With this in mind, we present the most prevalent (based on minor allele frequency) non-synonymous SNPs in the genomic regions of structural proteins. Our results indicate that there are distinct viral isolate clusters for SARS-CoV-2 sequences uploaded to GISAID. Larger outbreak clusters and hubs from UMAP and MST probably reflect outbreak epicentres (figure). Smaller clusters might be indicative of isolated outbreaks, and singleton samples might be indicative of isolated cases (figure). The analysis of SNPs reveals variants with notable minor allele frequencies involving missense substitutions in structural protein-coding genome regions (figure). Our novel tool is accessible to those who might not be trained in bioinformatics and epidemiological analysis, and thus it serves as a platform for aiding in a pivotal aspect of the global research effort against the COVID-19 pandemic. The current limitations of CGT include the sequence input limit and processing time, both of which are because of the size of the public data that must be concurrently processed with user-input sequences (>25 000 GISAID sequences). Our team is working to continuously integrate optimisations to the CGT data processing pipeline through code parallelisation and the refinement of deployment infrastructure. Future releases of the application will aim to decrease the processing time, increase the user-input sequence limit, incorporate the input and processing of raw SARS-CoV-2 sequencing data (eg, FASTQ files), and add additional information related to sequence epidemiology, such as travel history. Complete documentation of the CGT analysis pipeline and application source-code are available in a GitHub repository. Our application does not store any user-uploaded sequence data on the server-side or client-side; CGT simply processes the data to create updated visualisations, and information does not persist after the user disconnects. SARS-CoV-2 genome sequence data and linked metadata from GISAID are not published on our website, as per the GISAID data usage policy. Up-to-date acknowledgments for the usage of GISAID uploaded sequences for analysis are available in the GitHub repository. Our team is grateful to all the researchers who have shared SARS-CoV-2 viral genome sequencing data on GISAID. We declare no competing interests. HMa, BW, and AGM conceptualised and designed the study. HMa and HMb collected and analysed the data. HMa, HMb, BW, AB, JAN, ARR, and AGM interpreted the initial results. HMa developed the application. HMa and ARR did the software testing. HMa and NK created the application documentation. HMa, HMb, ARR, AB, JAN, RAK, NK, SM, AGM, and BW tested the application. RAK, NK, SM, and AGM provided suggestions for application visualisations. HMa and HMb created the manuscript figures. HMa and BW wrote the manuscript. HMa, BW, HMb, AB, NK, SM, and AGM edited the manuscript. eyJraWQiOiI4ZjUxYWNhY2IzYjhiNjNlNzFlYmIzYWFmYTU5NmZmYyIsImFsZyI6IlJTMjU2In0.eyJzdWIiOiJiNzZmZjEwZGVjYTlmNDkwYThmMzc1OTJiMzcwZWI4NCIsImtpZCI6IjhmNTFhY2FjYjNiOGI2M2U3MWViYjNhYWZhNTk2ZmZjIiwiZXhwIjoxNzAyMzgwMzczfQ.g2yjYNk7xVxCyBbm4vi8Aa-0h1-odsfYeHUd0iQQ8gnXPmhtlWvlHRLiY2uXC2fduG_XljWdVEN2XnbRGfXcAJxR8_45TAz2DwvjLWmQhtI7CS6Tapk7LPFaH4UI2EJsplugk6HSE1m9X_ez7hGNkQnBvVofeq7IaE3yZ0y_deIqIovCZtDulhoxi4UU11fDZXLkd0MyaibigCcGNZPFgG-Y6AyPRGqhkYQ0FwIgKoiuciUG89rJXdhquM-S7nipapeTGwhaJFdDjey5DDhU84kWbXNS83KoNiN7P-VWS2SAJu4hUx074AFB7djpl_m3JFzBhmBpJr27wdl831uYWw Download .mp4 (84.41 MB) Help with .mp4 files Supplementary VideoThe COVID-19 Genotyping Tool tutorialA short demonstration on using the COVID-19 Genotyping Tool application, understanding its features, and navigating the website Download .pdf (.55 MB) Help with pdf files Supplementary appendix 2