Research Projects

Viroporins are family of small, hydrophobic integral membrane proteins essential to the life cycle of diverse range of RNA and DNA viruses. While thousands of putative viroporin sequences are available in the standard databases like Uniprot, only a small number of them have been functionally characterized. These virus-coded channel proteins vary greatly in their structure and are known to perform multiple functions during the virus life cycle. M2 proton channel of influenza A virus is the prototype viroporin with provenance as an antiviral drug target. Literature shows rapid expansion of viroporin family to include significant human pathogens including SARS Cov-2, human immunodeficiency virus type 1 (HIV-1), picornaviruses, alphaviruses and paramyxovirus. As the diverse functions of viroporins continue to be identified, it has been shown that their function is essential to the virus life cycle making them potent drug targets. Structurally viroporins are 50-275 amino acids long and comprise of one, two or three potential trans-membrane domains. Small length requires them to oligomerize to form intact pore across the membrane. Examples range from tetrameric M2 proton channel (PDB ID 2RLF, 3BKD) from Influenza to heptameric channel from hepatitis C virus (PDB ID 2M6X). Recent advances in Natural Language Processing, word embedding have opened new possibilities for efficient extraction, characterization and integration of sequence and structure-based features of diverse viroporin sequences. Machine learning based protein redesign provides an attractive approach to address fundamental biological processes of viral infection and life cycle. Also, combining prediction models based on sequence and structure information provides the essential data for redesigning naturally occurring channel proteins as pore containing proteins.