Fig. 1

NERVE 2.0 working structure. Bacterial protein sequences are provided as an input FASTA proteome and undergo eight analytical steps: A Subcelloc predicts protein subcellular localization, B Adhesin returns the probability of a protein to be an adhesin, C Tmhelices predicts protein topology, D Loop Razor rescues membrane proteins reduced to their extracellular fragments, E Autoimmunity and Mouse Immunity which find respectively matches between the pathogen under analysis and human or mice proteomes F Conservation which detects conserved proteins between two input bacterial strains, G Virulent to infer presence of virulence factors and H Annotation to predict protein function. Then, the Select module I filters out PVCs, which meet specific requirements. Output results can be downloaded in .json, .csv, or.xlsx format. Epitope prediction J is performed after the Select module. Created with BioRender.com