What problem does it solve? Finding homologous protein sequences and inferring protein function from sequence similarity normally requires manual interaction with BLAST or MMseqs2 web portals, result parsing, and metric interpretation. This Skill automates the entire workflow from a raw amino acid sequence or FASTA file to a summarized Markdown report of top homologues. ## Core Features & Use Cases - Dual search methods: Runs fast MMseqs2 searches via the ColabFold API by default, with automatic fallback to comprehensive EBI BLAST when MMseqs2 fails or when BLAST is explicitly requested. - Validated database selection: Supports 25 curated UniProt, UniRef, and PDB database codes with strict validation against an allowed list. - Structured output: Generates Markdown tables of top hits with Q-Cov, E-value, sequence identity, and alignment length, plus JSON files for downstream programmatic analysis. - Use Case: A researcher has an uncharacterized protein sequence and wants to infer its function. The Skill searches for homologues, interprets alignment metrics, and summarizes the functions of the top 3-5 hits. ## Quick Start Find homologues for the protein sequence in my file query.fasta and summarize the likely function of the protein.
