File Collection

File Collection

File Collection is under Collect files. It provides database connectors and manual upload for files already on your computer.

Connectors

• RCSB PDB — experimentally determined structures.
• UniProt — protein records and sequences.
• AlphaFold DB — predicted protein structures, resolved from a UniProt accession or gene name.
• PubChem — compound records and chemical files.
• ChEMBL — compounds and measured bioactivity.
• InterPro — protein domains and family annotations.
• Add Evidence — add a file that is not available through a connector.

Each connector page documents its search fields and the file types it can provide. Scoring and Qualities are shown only where the connector has a meaningful scoring workflow — UniProt is the one here that does; the rest use the database's own ordering and do not ask for importance.

Shared search and save flow

1. Search using the connector's identifier or query.
2. Review the returned metadata and open Details for less common fields.
3. Click Add to shortlist on each result you want.
4. In the shortlist, choose any combination of file types for each item. Some useful types are selected by default; you can change them.
5. Once every result you want is on the shortlist, click Save shortlist.

Saving one shortlist item with several file types creates one evidence row per selected file. The rows share the same run, so they stay linked without forcing unrelated formats into one file.

The same entry and file type cannot be added twice to a shortlist. Selection clears after adding. Automatic fetching is attempted when the shortlist is saved; large or unavailable files remain in Missing Files, where file ↗ opens the source and Manual Upload attaches what you downloaded.

Papers

Most records cite the literature behind them — a structure's primary citation, a UniProt annotation's references. Where a connector captures those, a Papers dropdown appears on the shortlist row before you save. Tick any you want and they are saved as literature evidence alongside the record itself.

You can also do it afterwards: Group ▾ on a saved row lists the record's other file types and its citations, with Download for anything not yet saved and Show for anything already in the project.

A paper already saved anywhere in the project is recognised rather than saved a second time — the existing copy is simply linked to this record as well.

Results and navigation

All Files is a connector-scoped slice of the global Results Hub, not a separate data store. Results Hub organises saved material as Category → Tool → Run → Evidence. The run name records the search query and timestamp; each evidence row is named for the selected item and its file type.

Use Details for provenance, identifiers, links, validation information, and other metadata. The main result stays focused on the fields needed to choose an entry.