EMAP-SSN: An Embedding- and Multiple-Alignment-Integrated Sequence Similarity Network Platform for Interactive Exploration of Protein Sequence Space

Sequence similarity networks (SSNs) are graphical representations commonly used to study protein sequence relationships. Traditional SSN workflows often employ BLAST to compute sequence similarities, with external visualization tools generating the final networks. This separates raw sequence data from similarity calculations during visualization, complicating analyses that require detailed residue-level information.

To address this limitation, researchers have developed EMAP-SSN, an open-source, cross-platform software suite that integrates SSN computation, visualization, and analysis into streamlined workflows.

EMAP-SSN offers two primary approaches for sequence-similarity calculation: BLAST-based alignments and embedding-based alignments. The program directly connects network nodes to their original sequences and multiple alignments, enabling analyses at the residue level. Its modular architecture supports the addition of specialized functionalities and future development through command integration and browser-based utilities.

In a demonstration using fold-type IV pyridoxal 5'-phosphate-dependent enzymes, EMAP-SSN effectively links network topology with residue-level variation. This capability allows for the identification of sequence clusters, mapping of functional motifs, and detection of subgroup-specific conservation patterns. By bridging large protein sequence sets with experimentally verifiable hypotheses on enzyme function and potential targets for protein engineering, EMAP-SSN streamlines the process from protein sequence exploration to actionable biological insights. The EMAP-SSN software is available for download at https://github.com/Xuebin-Feng/EMAP-SSN.