Introduction
Identification of vaccine targets in pathogens using computational approaches
Identification of vaccine targets in pathogens using computational approaches
New data-driven approaches, such as reverse vaccinology, systems vaccinology, and machine learning, have started to capitalize on the vast potential of omics data available for vaccine design. Several computational studies have analyzed genomes or proteomes of individual pathogenic strains or species to find out vaccine candidates such as Goodswen et al., 2018 used a machine learning approach to distinguish between true and false vaccine candidates for eukaryotes including Caenorhabditis elegans, Toxoplasma gondii and Plasmodium sp.
Despite significant advancements in vaccinology, computational proteomics, machine learning, and reverse vaccinology domains, finding vaccine candidates, producing them in the laboratory, and confirming their efficacy in animal models remain complicated undertakings. To begin, there has been an urgent need for building pipelines or computational frameworks, to integrate diverse algorithms and databases using a single input and provide meaningful results for researchers working on vaccine development.
In this work, we are introducing an integrated framework that combines immuno-informatics approaches, bioinformatics tools, and supervised machine learning-based tools for vaccine discovery. Here, we attempt to rank or classify pathogen proteins based on their propensity to be good vaccine candidates and to design safe and effective multiple epitope vaccine candidates using a set of tools such as PsortB, BLAST, HMMTop, ProtParam, FungalRV, NetCTL, VaxiJen 2.0, IEDB tools, etc. As a proof of concept, we applied our system on different pathogens including Mycobacterium tuberculosis, Plasmodium vivax, Candida albicans, and Influenza A virus and identified several key vaccine candidates.