Prediction of Genomic Signature of Ngs Sequences and Comparative Drug-Likeness

Abstract

Developing a drug or particular immunotherapy medication for a worldwide epidemic illness caused by viruses (current pandemic) necessitates comprehensive evaluation and annotation of the metagenomic datasets to filter nucleotide sequences quickly and efficiently. Because of the homologs' origin of aligning sequences, space complexity, and time complexity of the analyzing system, traditional sequence alignment procedures are unsuccessful. This necessitates employing an alignment-free sequencing approach in this research that solves the foregoing issue. We suggest a distance function that compresses performance metrics for automatically identifying Short nucleotide sequences used by SARS coronavirus variants to identify critical features in genetic markers and genomic structure. This method provides easy recognition of data compressed by using a set of mathematical and computational tools in the study. We also show that by using our suggested technique to examine extremely short regions of nucleotide sequences, we can differentiate SAR-CoV-2 from SAR-CoV-1 viruses. Later, the Lipinski descriptor (rule of 5) was used to predict the drug-likeness of the target protein in SARS-CoV-2. A regression model using random forest was created to validate the machine learning model for computational analysis. This work was furthered by comparing the regressor model to other machine learning models using lezypredict, allowing scientists to swiftly and accurately identify and describe the SARS coronavirus strains.