Variant analysis of SARS-CoV-2 genomes in the Middle East

Abstract

Background Coronavirus (COVID-19) was introduced into society in late 2019 and has now reached over 88 million cases and 1.9 million deaths. The Middle East has a death toll of ~80,000 and over 35000 of these are in Iran, which has over 1.2 million confirmed cases. We expect that Iranian cases caused outbreaks in the neighbouring countries and that variant mapping and phylogenetic analysis can be used to prove this. We also aim to analyse the variants of severe acute respiratory syndrome coronavirus-2 (SARS -CoV-2) to characterise the common genome variants and provide useful data in the global effort to prevent further spread of COVID-19. Methods The approach uses bioinformatics approaches including multiple sequence alignment, variant calling and annotation and phylogenetic analysis to identify the genomic variants found in the region. The approach uses 122 samples from the 13 countries of the Middle East sourced from the Global Initiative on Sharing All Influenza Data (GISAID). Findings We identified 2200 distinct genome variants including 129 downstream gene variants, 298 frame shift variants, 789 missense variants, 1 start lost, 13 start gained, 1 stop lost, 249 synonymous variants and 720 upstream gene variants. The most common, high impact variants were 10818delTinsG, 2772delCinsC, 14159delCinsC and 2789delAinsA. These high impact variant ultimately results in 36 number of mutations on spike glycoprotein. Variant alignment and phylogenetic tree generation indicates that samples from Iran likely introduced COVID-19 to the rest of the Middle East. Interpretation The phylogenetic and variant analysis provides unique insight into mutation types in genomes. Initial introduction of COVID-19 was most likely due to Iranian transmission. Some countries show evidence of novel mutations and unique strains. Increased time in small populations is likely to contribute to more unique genomes. This study provides more in depth analysis of the variants affecting in the region than any other study.

Tracking SARS-CoV-2 Spike Protein Mutations in the United States (2020/01 – 2021/03) Using a Statistical Learning Strategy

“…The NextStrain methodology (18) can model dynamic changes of variant proportions, while an alternative approach aligns sequence data to a matrix of binary indicators for the presence of variants, and systematically evaluates each mutant as a potential variant (19). Leveraging the analytic approach of single nucleotide polymorphisms (SNPs), variants have been identified by assessing linkage-disequilibrium (20) or similar SNPbased identification and analysis (21). However, with the exception of the NextStrain methodology ( 3), these methods do not directly take into account sequence collection time, nor explicitly incorporate highly granular geographic information.…”

SARS-CoV-2 surveillance in Italy through phylogenomic inferences based on Hamming distances derived from functional annotations of SNPs, MNPs and InDels

“…Then, these de novo assemblies and consensus sequences are usually aligned between each other through multiple sequence alignment (MSA) (e.g. Muscle [27], Augur toolkit [31], MAFFT [32], Clustal (O/W) [33]) in order to perform substitution model-based phylogenomic inferences through maximum likelihood (ML) (e.g. IQ-TREE [31], RaxML [27]) or Bayesian models (BEAST [33]).…”

“…Muscle [27], Augur toolkit [31], MAFFT [32], Clustal (O/W) [33]) in order to perform substitution model-based phylogenomic inferences through maximum likelihood (ML) (e.g. IQ-TREE [31], RaxML [27]) or Bayesian models (BEAST [33]). The aligned de novo assemblies and consensus sequences can also be derived (i.e.…”

“…snp-sites) into variant calling format (i.e. VCF) [33]. Because the biological effects of variants (i.e.…”

“…SNPs, MNPs and InDels, so-called genotypes in VCF files) are required for identifying SARS-CoV-2 lineages (e.g. PANGOLIN [34]) and accordingly designing of vaccines [35], these VCF files from variant calling or aligned sequences are usually used as input of functional annotation of variants with SNPeff [27, 33] or ANNOVAR [28]. While de novo assembly (i.e.…”