Determination of secretion type associated motifs from pathogenic protein sequences

Abstract

Currently, potential vaccine candidates are determined using protein localization predictors to identify surface or secreted protein sequences. The pathogenic bacterial sequences constitute only a tiny percentage of this database. They search for general motifs found in bacteria and miss out on pathogen-specific motifs, which is crucial for identifying vaccine candidates. We developed a motif search algorithm derived from pathogenic bacterial sequences in this work. Using this method, we identified secretion motifs of pathogenic bacteria for six different secretion pathways and non-specific secretion motifs. To this end, we established a database of immunogenic sequences from patented vaccine sequences, AntigenDb, and PUB-MED. The motifs are filtered out according to their presence in a cytoplasmic dataset to keep motifs that are present only in secreted sequences. These sequences are validated on newly identified novel vaccine candidates from 20 pathogenic bacterial immunoproteomics data.