Background: Short Linear Motifs (SLiMs) are short (3-15aa) segments of proteins that mediate numerous protein-protein interactions (PPI) in critical biological pathways and signaling networks. SLiMs typically occur in structurally disordered regions of proteins and have few sites specific to function, resulting in high evolutionary plasticity and frequent convergent evolution on different protein backgrounds.  Such “molecular mimicry” is abundant in viruses, which exploit SLiMs to influence the molecular machinery of host cells.

Methods: We are combining publicly available datasets of host-host and host-pathogen PPI with recently developed tools from the SliMSuite package (1) SLiMProb to identify novel candidates for viral mimicry of known host SLiMs [1], and (2) QSLiMFinder to predict entirely new SLiM classes [2]. In addition to novel SLiMs, we also aim to predict new human targets for known viral SLiMs. In each case, signals of convergent evolution are identified using statistical over-representation of motifs in unrelated proteins. SLiM predictions will be put in context using network analysis of the host interactome.

Results: The approach will be benchmarked using known cases of molecular mimicry in viral proteins. Novel candidates of molecular mimicry that appear to target important proteins and pathways along with GO annotations for the viral life cycle will be highlighted.

Conclusion: Systems-level analysis of molecular mimicry in virus-host and host-host PPI data using high throughput computational SLiM discovery tools has great potential to increase our understanding of how viruses manipulate their hosts and identify candidates for novel therapeutic targets. Although still in its early stages, this work reveals key considerations for future analysis.