Automatic GO Annotation of Gene Products in SARS-CoV-2

The main goal of SARS-Cov2 gene product annotation is to identify the protein’s role in biological processes and disease-leading genes. Protein annotation involves a combination of experimental and computational approaches. The computational one guides the potential annotation and the experimental sustain it. Herein, we tackle a computational analysis by a machine learning tool carefully tailored to virus annotations, called FGGA, which can predict annotations across all three Gene Ontology (GO) subdomains. With this goal, we selected 3 proteins associated with structural, non-structural, and accessory SARS-Cov2 proteins. We analyzed the results by comparing the predicted GO terms and well-registered annotations at the Uniprot database. We found that FGGA can guide experts with potential roles of viral adaptation and pathogenicity, if proper characterization of a representative number of viral sequences is available.