Improvements for Unique Peptide Finder support

[pverscha]

I've recently been working on a new prototype of a Unique Peptide Finder tool. This tool is still under active development, but will be a good addition to the Unipept Web application when it's ready.

In order for this new tool to work efficiently, I've had to implement a few new endpoints for the API and had to make some changes to pre-existing endpoints. In addition to new functionality, I was also able to drastically speed up the pept2taxa and pept2lca endpoints. This PR bundles all these changes.

Important

This PR also required changes to the unipept-index package, which have been published in this PR. Make sure to merge that PR first, before deploying this one. Note that we also need to remove the branch = reference from the Cargo.toml file in the index package reference.

New endpoints added

private_api/taxa/unique_peptides

• Adds a new GET/POST endpoint at /private_api/taxa/unique_peptides (and .json) that computes taxon-unique peptides for a given strain or species.
• Adds get_proteins_for_taxon() to the database crate: an exact term-query on taxon_id with search_after pagination to correctly handle taxa with more than 10,000 proteins.
• Adds fancy-regex dependency to support lookahead assertions in user-supplied cleavage patterns (the standard regex crate does not support lookaheads).

Request parameters

Parameter	Type	Required	Default	Description
taxon_id	u32	yes	—	NCBI taxon ID; must be at species or strain rank
cleavage_regex	string	no	[KR](?!P)	Regex matching cleavage sites; split occurs after each match (standard tryptic convention)
min_length	usize	no	5	Minimum peptide length in amino acids

Example response

{
  "unique_peptides": ["AAFEDLQSLQDK", "NLFVAKNLR"],
  "total_peptides": 4821,
  "total_unique_peptides": 312
}

total_peptides is the count of deduplicated peptides after digestion and length filtering. total_unique_peptides equals unique_peptides.length.

Error cases:

• 400 — invalid cleavage_regex
• 400 — taxon_id not found in the taxon store
• 400 — taxon_id is not at species or strain rank (message includes the actual rank)

private_api/taxa/shared_peptides

• Adds a new GET/POST endpoint at /private_api/taxa/shared_peptides (and .json) that computes peptides shared across all provided taxa.
• Reuses get_proteins_for_taxon() from the database crate (added for the unique_peptides endpoint) and the shared digestion helpers in api/src/helpers/digestion.rs.

Request parameters

Parameter	Type	Required	Default	Description
taxon_ids	u32[]	yes	—	List of NCBI taxon IDs; each must be at species or strain rank
cleavage_regex	string	no	[KR](?!P)	Regex matching cleavage sites; split occurs after each match (standard tryptic convention)
min_length	usize	no	5	Minimum peptide length in amino acids

Example response

{
  "shared_peptides": ["AAFEDLQSLQDK", "NLFVAKNLR"]
}

shared_peptides contains deduplicated peptides (after digestion and length filtering) that are present in at least one protein of every provided taxon. The list is sorted lexicographically. If taxon_ids is empty, an empty list is returned.

Error cases:

• 400 — invalid cleavage_regex
• 400 — any taxon_id not found in the taxon store
• 400 — any taxon_id is not at species or strain rank (message includes the actual rank)

Changes made to existing endpoints

private_api/taxa

Adds a new parameter report_protein_count that will report the amount of proteins that are associated with the provided taxa. If a taxon is provided at a level that's higher than species or strain, the protein count is equal to the sum of the protein count of all descendant taxa.

Example request:

{
    "taxids": [4751],
    "report_protein_count": true
}

Example response:

[
    {
        "id": 4751,
        "name": "Fungi",
        "rank": "kingdom",
        "lineage": [
            2759,
            null,
            4751,
            null,
            null,
            null,
            null,
            null,
            null,
            null,
            null,
            null,
            null,
            null,
            null,
            null,
            null,
            null,
            null,
            null,
            null,
            null,
            null,
            null,
            null,
            null,
            null,
            null
        ],
        "protein_count": 38050
    }
]

api/v2/pept2taxa

Improved performance by switching to the more lightweight analyse_taxa index function instead of analyse (see PR: unipept/unipept-index#36)

api/v2/pept2lca

Improved performance by switching to the more lightweight analyse_taxa index function instead of analyse (see PR: unipept/unipept-index#36)