Rapid in silico design of antibodies targeting SARS-CoV-2 using machine learning and supercomputing

Thomas Desautels; Adam Zemla; Edmond Lau; Magdalena Franco; Daniel Faissol

In the absence of a known SARS-CoV-2 spike protein structure, we characterized the surface glycoprotein sequence YP_009724390.1 [13] by constructing a homology-based structural model using the AS2TS system [14] and SARS-CoV-1. To assess regions of sequencestructure conservation or variability between SARS-CoV-1 and SARS-CoV-2 (sequence similarity=71%), we created a set of initial structural models using available templates. The comparative analysis between constructed models and all available templates from PDB was performed using the StralSV algorithm [28] , which identifies protein fragments that exhibit structural similarities despite low primary amino acid sequence similarity and regions where some structure conformation uncertainties can be observed. We used results from these searches for modeling missing loop regions and assessment of possible structure conformation diversity in preliminary models.