find_linear_motifs
Read-only
Short linear motifs in a protein: degrons, docking sites, targeting signals, cleavage and modification sites — the three-to-eleven-residue functional sites a profile-HMM domain scan cannot see.
Input Schema
| Name | Required | Description | Default |
|---|---|---|---|
| _files | No | The project: record references shaped like _subject. Their records become library documents a tool names in sequences/parts/document; a trace or alignment is found by file name. | |
| domains | No | the globular domains to bury motifs in, as `start-end` residue ranges (1-based, inclusive), comma-separated, each optionally labeled `start-end@label` so a dropped hit can say what buried it — e.g. `20-153@G3DSA:3.90.180.10, 296-341@G3DSA:3.90.180.10`. Pass the FRAGMENTS of an InterProScan domain match, not its outer span. Empty falls back to the scan the Protein domains panel holds, which is the only source the in-browser doors need. Only applies when filterInDomains is true. | |
| protein | Yes | the protein sequence, one-letter codes; wire it from translate_sequence or read it off a CDS. | |
| species | No | the species the protein comes from, as an NCBI Taxonomy id (e.g. 4932) or a scientific name (e.g. Saccharomyces cerevisiae), for ELM's taxonomic range filter: a motif whose class ELM does not place in that species' lineage is dropped, with its range. Empty runs no taxon filter, which is ELM's own default. | |
| _subject | No | The record to act on: exactly one of sequence, accession (+source), sample (an id from list_samples) or file ({name, text} or {name, base64}); record_index picks one record of several. | |
| families | No | restrict to these ELM families, comma-separated, from CLV, DEG, DOC, LIG, MOD, TRG. Empty is no restriction and scans every family; the literal "none" is a restriction to no family and scans nothing, which is what the panel does when every family is unticked. | |
| compartments | No | restrict to motifs whose class is annotated for these cell compartments, as GO ids, comma-separated (e.g. GO:0005634 for nucleus). Empty runs no compartment filter, which is ELM's own default. | |
| expectCutoff | No | ELM's class gate: classes whose chance probability per residue exceeds this are not scanned at all. ELM's own default is 0.01; 0 scans every class, which is the resource's “all found (before filtering)” listing. | |
| filterInDomains | No | apply ELM's globular-domain filter, dropping motifs buried inside a folded domain. Takes the domains from `domains` when you pass them, and otherwise from the scan the Protein domains panel currently holds; with neither, the call refuses rather than silently not filtering. | |
| structuralContext | No | score each motif's structural context, as ELM's structure filter does. Empty applies nothing, which is the panel's own default. `alphafold` uses the predicted model and yields Q_acc only. A four-character PDB id (e.g. `1crn`) uses that experimental entry and yields Q_acc, Q_sse and ELM's three-way verdict, because only a real entry has a DSSP file. Either way the protein is resolved to its UniProt entry by sequence checksum first, and the call refuses if it is not an exact match. |
Output Schema
| Name | Required | Description | Default |
|---|---|---|---|
| hits | Yes | each surviving motif, in the engine’s own field names: identifier, accession, functionalSiteName, description, 1-based start/end, the residues that matched, the class chance probability, and how many experimentally annotated instances ELM holds | |
| count | Yes | surviving motifs, after every filter that ran | |
| accounting | Yes | what the scan was actually run against — classes scanned, classes skipped by the cutoff, classes skipped by the family selection — so a hit count can be read against its denominator rather than on its own | |
| provenance | Yes | the ELM release and date the scan ran against, and the filters that ran |