Figure 3.

In silico prediction of M. tuberculosis adhesins. (A) Various in silico prediction methods for identifying M. tuberculosis adhesins described in the literature. (B) Bioinformatic pipeline for predicting adhesin-like proteins from M. tuberculosis using latest general prediction methods. This pipeline consists of five sequential stages. In Stage 1, the complete proteome of M. tuberculosis H37Rv is retrieved and screened using the SPAAN to predict potential adhesins based on amino acid composition and related features. In Stage 2, functionally annotated proteins are selected using Mycobrowser, focusing on categories relevant to host–pathogen interaction. Stage 3 refines the candidates by identifying conserved hypothetical proteins, defined as uncharacterized proteins conserved across mycobacterial species, while excluding known proteins using multiple databases. In Stage 4, structure prediction is performed using Phyre2, and functional domains are analyzed using Pfam, complemented by the literature-based functional assessment. Stage 5 involves subcellular localization and secretion prediction using tools such as SubLoc, PsortB, and SignalP, prioritizing extracellular and membrane-bound proteins. Each of the five steps is visually represented as a separate module in the diagram, with arrows indicating the flow of analysis.