mimicINT

mimicINT infers large-scale protein-protein interactions between microbes and human hosts by detecting putative molecular mimicry elements, including short linear motifs and host-like globular domains, that mediate interactions with host proteins.

Key Features:

• Molecular mimicry detection: Identifies short linear motifs (SLiMs) and host-like globular domains in microbial proteins that potentially mediate interactions with human proteins.
• Template-based interaction inference: Predicts microbe-host protein interactions using known templates of domain-domain and SLiM-domain interactions.
• Monte-Carlo statistical assessment: Implements robust Monte-Carlo simulations to assess the statistical significance of SLiM detection and control for false positives.
• Interaction specificity filter: Accounts for differences between motif-binding domains of the same family to improve interaction prediction specificity.

Scientific Applications:

• Pathogenic effector interface identification: Identifies potential interfaces in experimentally detected interactions between pathogenic Escherichia coli type-3 secreted effectors and human proteins.
• Viral-host interaction inference: Infers biologically relevant interactions between Marburg virus and human proteins.
• Microbial infection mechanism characterization: Assists in characterizing the molecular mechanisms of microbial infections by predicting microbe-host protein-protein interactions.

Methodology:

mimicINT detects putative molecular mimicry elements (short linear motifs and host-like globular domains) in microbial proteins and exploits these elements to infer interactions with human proteins using known domain-domain and SLiM-domain interaction templates. Statistical significance of SLiM detection is assessed via Monte-Carlo simulations, and an interaction specificity filter accounts for differences between motif-binding domains of the same family.