AIVE: accurate predictions of SARS-CoV-2 infectivity from comprehensive analysis

Abstract

AbstractThis study presents an innovative research model utilizing big data science and protein structure prediction AI software. An unprecedented amount of SARS-CoV-2 data has been accumulated compared with previous infectious diseases, enabling insights into its evolutionary process and more thorough analyses. We identified amino acid substitutions ranging from hydrophilic to hydrophobic, or positively charged amino acids in the RBM region. An increased frequency of amino acid substitutions to lysine (K) and arginine (R) was detected in Variants of Concern (VOCs) and viral sequencing data. As the virus evolved to Omicron, commonly occurring mutations became fixed components of the new viral sequence. Furthermore, in specific positions, only one type of amino acid substitution and a notable absence of mutations at D467 was detected across viral sequences in VOCs. The binding affinity with the ACE2 receptor increased for later lineages. We developed APESS, a mathematical model evaluating infectivity based on biochemical and mutational properties calculated from a protein prediction of AlphaFold. We validated discoveries of features found through APESS. Infectivity was evaluated in silico using real-world viral sequences and in vitro viral entry assays. Using Machine Learning, we predicted mutations that had the potential to become more prominent. APESS and characteristics we discovered are featured in AIVE, a web-based system, accessible athttps://ai-ve.org. AIVE provides an infectivity measurement of mutations entered by users which is available on fast APESS calculations and visualization of results without GPU installation. We established a clear link between specific viral properties and increased infectivity. Comprehensive analysis and specialized AIVE reporting enhance our understanding of SARS-CoV-2 and enable more accurate predictions of infectivity.