Prediction of virulence factors in bacterial hypothetical proteins

A large number of genes in bacterial genomes encode hypothetical proteins (HPs) with unknown functions that require annotation. These HPs have recently been linked to pathogenicity due to potential virulence factors. Understanding the key elements influencing their virulence is a crucial goal. In our study, we analyzed protein domain families from the Pfam database in select Gram-positive and Gram-negative bacteria to identify virulent factors using a scoring system. We employed computational tools such as VICMpred and VirulentPred to identify virulent domains in different bacteria and used regression and Support Vector Machine (SVM) with Weka to find the best candidates. Genotypes with higher adherence suggest that diverse adhesin expression contributes to varying virulence potential, making strains with higher adhesin expression more aggressive. We believe our methods for predicting bacterial protein functions can aid in developing drugs and vaccines by enabling the prediction of virulence proteins.