VPAgs-Dataset4ML: A Dataset to Predict Viral Protective Antigens for Machine Learning-Based Reverse Vaccinology

This dataset contains 2,145 viral protein sequences, with positive.fasta containing 210 protective antigens (PAgs) and negative.fasta containing 1,935 non-protective protein sequences. The PAgs are the result of quality checks performed on viral PAgs from the publicly available Protegen database. The non-protective protein sequences were chosen using computational steps carried out on viral protein sequences in UniProt, a well-known and freely accessible protein sequence database. Researchers may benefit from VPAgs-Dataset4ML in developing machine learning models to predict viral protective antigens as potential vaccine candidates. This information could be used by vaccinologists to help develop effective vaccines that could help save patients' lives.