A Machine Learning approach for pre-miRNA discovery in SARS-CoV-2

G. MERINO, J. RAAD, L. BUGNON, D.H. MILONE, G. STEGMAYER; F. ARIEL

virtual

Conferencia; Intelligent Systems for Molecular Biology (ISMB) 2021; 2021

International Society for Computational Biology (ISCB)

Motivation The Severe AcuteRespiratory Syndrome-Coronavirus 2 (SARS-CoV-2) is responsible for the pandemicoutbreak of the coronavirus disease (COVID-19). MiRNA-like sequences have beenidentified in genomes of RNA viruses that replicate in the cytoplasm. Even more,it has been recently found that the SARS-CoV-1 encodes small viral RNAsfunctionally linked to the related lung pathogenesis. However, the capacity ofSARS-CoV-2 to encode functional putative microRNAs (miRNAs) remains largelyunexplored.Results We have developed anovel approach based on machine learning for identifying precursors ofmicroRNAs (pre-miRNAs) in the genome of the novel coronavirus [1]. Theseprecursors are processed by the cell to obtain miRNAs: a special type of smallnon-coding RNA of about 22 nucleotides that participate in gene regulationinfluencing diverse biological processes such as development, proliferation anddifferentiation across different cell types [2], with important roles indisease development and progression. The quantification of the abundance ofspecific miRNAs can assist in diagnosis, prognosis prediction and therapeuticassessment for various diseases including viral infections [3]. Remarkably,host miRNAs have been recently associated with antiviral defense mechanismstriggered by a coronavirus [4], and the activity of miRNAs derived from viralgenomes has also been proved lately [5,6]. The discovery of miRNAs in the novelvirus is of high importance in the context of the current worldwide sanitarycrisis, especially for contributing to the improvement of diagnostic andtreatment strategies [7].