# STRING

> Explore protein-protein interactions with STRING, the database of known and predicted protein networks for over 12,000 organisms: map gene names to proteins, get interaction networks and partners with confidence scores, run functional enrichment (GO, KEGG, Reactome) and interaction enrichment on a gene list, find homologs, and link to network pictures.

App · 13 actions. Web page: https://docs.lodol.com/docs/integrations/string-db

## Actions

### Describe a gene set

Get up to three short descriptions of what a set of proteins is about, chosen by STRING from its enriched terms, such as 'Cellular response to gamma radiation'.

**Inputs**

- Proteins (`identifiers`, list, required): The proteins, as gene names (TP53), UniProt accessions (P04637) or STRING IDs (9606.ENSP00000269305), one per line. STRING IDs from 'Map identifiers to STRING' are fastest.
- Organism (`species`, text, required): The proteins' organism. Pick one, or type its NCBI taxon ID (9606 for human) or name. 'Find species' looks up any other.

**Outputs**

- `descriptions` (list): The descriptions found, best first.
- `found` (true/false): False when STRING found nothing that fits the set.
- `primary_description` (text): The best few words for the set.
- `secondary_description` (text): The next best, when there is one.
- `summary` (text): One line saying what came back.
- `tertiary_description` (text): The third, when there is one.

**Example output**

```json
{
  "descriptions": [
    "Cellular response to gamma radiation",
    "Platinum drug resistance",
    "p53-Dependent G1 DNA Damage Response"
  ],
  "found": true,
  "primary_description": "Cellular response to gamma radiation",
  "secondary_description": "Platinum drug resistance",
  "summary": "Cellular response to gamma radiation; Platinum drug resistance; p53-Dependent G1 DNA Damage Response",
  "tertiary_description": "p53-Dependent G1 DNA Damage Response"
}
```

### Find closest homologs in other organisms

Find each protein's most similar protein in other organisms, such as the mouse counterpart of a human gene, with its name, STRING ID and bit score.

**Inputs**

- Proteins (`identifiers`, list, required): The proteins, as gene names (TP53), UniProt accessions (P04637) or STRING IDs (9606.ENSP00000269305), one per line. STRING IDs from 'Map identifiers to STRING' are fastest.
- Organism (`species`, text, required): The proteins' organism. Pick one, or type its NCBI taxon ID (9606 for human) or name. 'Find species' looks up any other.
- Other organisms (`target_species`, text, required): Where to look. Pick one, or type NCBI taxon IDs separated by commas (10090, 7227).

**Outputs**

- `count` (number): How many were found.
- `homologs` (list): One closest protein per protein and organism.
- `summary` (text): One line saying what came back.

**Example output**

```json
{
  "count": 2,
  "homologs": [
    {
      "bitscore": 598.6,
      "homolog_name": "Cdk1",
      "homolog_species": "Mus musculus",
      "homolog_string_id": "10090.ENSMUSP00000020099",
      "homolog_taxon_id": "10090",
      "protein": "CDK1",
      "string_id": "9606.ENSP00000378699",
      "taxon_id": "9606"
    },
    {
      "bitscore": 460.7,
      "homolog_name": "Cdk1",
      "homolog_species": "Drosophila melanogaster",
      "homolog_string_id": "7227.FBpp0079641",
      "homolog_taxon_id": "7227",
      "protein": "CDK1",
      "string_id": "9606.ENSP00000378699",
      "taxon_id": "9606"
    }
  ],
  "summary": "2 closest homologs: Cdk1 (Mus musculus), Cdk1 (Drosophila melanogaster)"
}
```

### Find proteins for a term

Find the proteins an organism has for a pathway, process, disease, tissue or domain, by the term's ID (hsa04115, GO:0006915) or by words in its name (melanoma). Best match first; each term lists up to about 100 proteins.

**Inputs**

- Term (`term`, text, required): A term ID, or words from its name.
- Organism (`species`, text, required): The proteins' organism. Pick one, or type its NCBI taxon ID (9606 for human) or name. 'Find species' looks up any other.

**Outputs**

- `count` (number): How many terms match.
- `found` (true/false): False when no term matches.
- `summary` (text): One line saying what came back.
- `terms` (list): The matching terms, best match first.

**Example output**

```json
{
  "count": 1,
  "found": true,
  "summary": "1 term matches 'hsa04115'; best: p53 signaling pathway (hsa04115), 75 proteins",
  "terms": [
    {
      "category": "KEGG",
      "category_name": "KEGG pathways",
      "description": "p53 signaling pathway",
      "protein_count": 75,
      "proteins": [
        "GADD45B",
        "TSC2",
        "TNFRSF10A",
        "SERPINE1",
        "CD82",
        "CCND1"
      ],
      "string_ids": [
        "9606.ENSP00000215631",
        "9606.ENSP00000219476",
        "9606.ENSP00000221132",
        "9606.ENSP00000223095",
        "9606.ENSP00000227155",
        "9606.ENSP00000227507"
      ],
      "term": "hsa04115"
    }
  ]
}
```

### Find species

Find organisms in STRING by name or taxon ID, to get the taxon ID the other steps take. Core organisms, whose networks draw on every kind of evidence, come first.

**Inputs**

- Name or taxon ID (`query`, text, required): Part of an organism's name, or its NCBI taxon ID.
- Core organisms only (`core_only`, true/false): Only STRING's core organisms, whose networks draw on every kind of evidence.
- Most organisms (`limit`, number): The most organisms to return, 1 to 100. Defaults to 20. From 1 to 100.

**Outputs**

- `count` (number): How many are listed.
- `found` (true/false): False when no organism matches.
- `organisms` (list): The matching organisms, best match first.
- `summary` (text): One line saying what came back.
- `total` (number): How many organisms match in all.

**Example output**

```json
{
  "count": 2,
  "found": true,
  "organisms": [
    {
      "domain": "Bacteria",
      "is_core": true,
      "name": "Escherichia coli K12",
      "official_name": "Escherichia coli str. K-12 substr. MG1655",
      "taxon_id": "511145"
    },
    {
      "domain": "Bacteria",
      "is_core": false,
      "name": "Escherichia coli CFT073",
      "official_name": "Escherichia coli CFT073",
      "taxon_id": "199310"
    }
  ],
  "summary": "5 organisms match 'Escherichia coli'; best: Escherichia coli K12 (taxon 511145)",
  "total": 5
}
```

### Get functional annotation

Get every term STRING assigns to a set of proteins (GO terms, pathways, keywords, domains, tissues, diseases...), whether or not it is enriched, with the proteins carrying each. Papers are included only when their category is picked.

**Inputs**

- Proteins (`identifiers`, list, required): The proteins, as gene names (TP53), UniProt accessions (P04637) or STRING IDs (9606.ENSP00000269305), one per line. STRING IDs from 'Map identifiers to STRING' are fastest.
- Organism (`species`, text, required): The proteins' organism. Pick one, or type its NCBI taxon ID (9606 for human) or name. 'Find species' looks up any other.
- Categories (`categories`, list): Only terms from these categories. Leave empty for all of them. One or more of: `Process` (GO biological process), `Function` (GO molecular function), `Component` (GO cellular component), `KEGG` (KEGG pathways), `RCTM` (Reactome pathways), `WikiPathways`, `Hallmark` (Hallmark gene sets (MSigDB)), `Keyword` (UniProt keywords), `Pfam` (Pfam domains), `InterPro` (InterPro domains and features), and 13 more.
- Most terms (`limit`, number): The most terms to return, 1 to 2,000. Defaults to 200. From 1 to 2000.

**Outputs**

- `annotations` (list): The terms, grouped by category.
- `count` (number): How many terms are listed.
- `summary` (text): One line saying what came back.
- `total` (number): How many terms there were before the limit.

**Example output**

```json
{
  "annotations": [
    {
      "category": "KEGG",
      "category_name": "KEGG pathways",
      "description": "Endocrine resistance",
      "gene_count": 2,
      "genes": [
        "MDM2",
        "TP53"
      ],
      "ratio_in_set": 1.0,
      "string_ids": [
        "9606.ENSP00000258149",
        "9606.ENSP00000269305"
      ],
      "term": "hsa01522"
    },
    {
      "category": "KEGG",
      "category_name": "KEGG pathways",
      "description": "Platinum drug resistance",
      "gene_count": 2,
      "genes": [
        "MDM2",
        "TP53"
      ],
      "ratio_in_set": 1.0,
      "string_ids": [
        "9606.ENSP00000258149",
        "9606.ENSP00000269305"
      ],
      "term": "hsa01524"
    }
  ],
  "count": 2,
  "summary": "2 annotations for 2 proteins",
  "total": 2
}
```

### Get interaction network

Get the interactions among a set of proteins, each with STRING's confidence score and the evidence behind it (experiments, databases, text mining, co-expression...). Choose functional, physical or regulatory interactions and a minimum confidence, and optionally widen the network with the closest partners.

**Inputs**

- Proteins (`identifiers`, list, required): The proteins, as gene names (TP53), UniProt accessions (P04637) or STRING IDs (9606.ENSP00000269305), one per line. STRING IDs from 'Map identifiers to STRING' are fastest.
- Organism (`species`, text, required): The proteins' organism. Pick one, or type its NCBI taxon ID (9606 for human) or name. 'Find species' looks up any other.
- Minimum confidence (`required_score`, text): How sure STRING must be of an interaction to include it. Leave empty for STRING's default, usually medium. One of: `900` (Highest confidence (0.900)), `700` (High confidence (0.700)), `400` (Medium confidence (0.400)), `150` (Low confidence (0.150)).
- Network type (`network_type`, text): Functional takes every kind of evidence that proteins work together; physical, only that they bind; regulatory, which one regulates the other. Defaults to functional. One of: `functional` (Functional associations (all evidence)), `physical` (Physical interactions (proteins that bind)), `regulatory` (Regulatory (who regulates whom)). Default: `functional`.
- Extra proteins (`add_nodes`, number): Widen the network with this many of the proteins' closest partners, 0 to 500. With a single protein, STRING adds 10 unless told otherwise. From 0 to 500.

**Outputs**

- `count` (number): How many interactions were found.
- `interactions` (list): The interactions found.
- `network_type` (text): functional, physical or regulatory.
- `protein_count` (number): How many proteins are in the network.
- `proteins` (list): Every protein in the network, by name.
- `required_score` (number): The minimum score asked for, 0 to 1000; empty for STRING's default.
- `summary` (text): One line saying what came back.
- `taxon_id` (text): The NCBI taxon ID of the organism; empty with no interactions.

**Example output**

```json
{
  "count": 3,
  "interactions": [
    {
      "coexpression_score": 0.097,
      "cooccurrence_score": 0.0,
      "database_score": 0.75,
      "directed": false,
      "experimental_score": 0.902,
      "fusion_score": 0.0,
      "neighborhood_score": 0.0,
      "protein_a": "MDM2",
      "protein_b": "ATM",
      "score": 0.995,
      "sign": null,
      "string_id_a": "9606.ENSP00000258149",
      "string_id_b": "9606.ENSP00000278616",
      "textmining_score": 0.835
    },
    {
      "coexpression_score": 0.129,
      "cooccurrence_score": 0.0,
      "database_score": 0.9,
      "directed": false,
      "experimental_score": 0.999,
      "fusion_score": 0.0,
      "neighborhood_score": 0.0,
      "protein_a": "MDM2",
      "protein_b": "TP53",
      "score": 0.999,
      "sign": null,
      "string_id_a": "9606.ENSP00000258149",
      "string_id_b": "9606.ENSP00000269305",
      "textmining_score": 0.984
    },
    {
      "coexpression_score": 0.075,
      "cooccurrence_score": 0.0,
      "database_score": 0.0,
      "directed": false,
      "experimental_score": 0.999,
      "fusion_score": 0.0,
      "neighborhood_score": 0.0,
      "protein_a": "MDM2",
      "protein_b": "CDKN1A",
      "score": 0.999,
      "sign": null,
      "string_id_a": "9606.ENSP00000258149",
      "string_id_b": "9606.ENSP00000384849",
      "textmining_score": 0.855
    }
  ],
  "network_type": "functional",
  "protein_count": 4,
  "proteins": [
    "MDM2",
    "ATM",
    "TP53",
    "CDKN1A"
  ],
  "required_score": 400,
  "summary": "3 functional interactions among 4 proteins (MDM2, ATM, TP53, CDKN1A)",
  "taxon_id": "9606"
}
```

### Get interaction partners

Get the proteins that interact with each of a set, from the whole of the organism's proteins, most confident first, with the score and evidence for each.

**Inputs**

- Proteins (`identifiers`, list, required): The proteins, as gene names (TP53), UniProt accessions (P04637) or STRING IDs (9606.ENSP00000269305), one per line. STRING IDs from 'Map identifiers to STRING' are fastest.
- Organism (`species`, text, required): The proteins' organism. Pick one, or type its NCBI taxon ID (9606 for human) or name. 'Find species' looks up any other.
- Minimum confidence (`required_score`, text): How sure STRING must be of an interaction to include it. Leave empty for STRING's default, usually medium. One of: `900` (Highest confidence (0.900)), `700` (High confidence (0.700)), `400` (Medium confidence (0.400)), `150` (Low confidence (0.150)).
- Network type (`network_type`, text): Functional takes every kind of evidence that proteins work together; physical, only that they bind; regulatory, which one regulates the other. Defaults to functional. One of: `functional` (Functional associations (all evidence)), `physical` (Physical interactions (proteins that bind)), `regulatory` (Regulatory (who regulates whom)). Default: `functional`.
- Partners per protein (`limit`, number): The most partners to return for each protein, 1 to 500. Defaults to 10. From 1 to 500.

**Outputs**

- `count` (number): How many interactions were found.
- `interactions` (list): Each protein's interactions with its partners, best first.
- `network_type` (text): functional, physical or regulatory.
- `partner_count` (number): How many different partners were found.
- `partners` (list): The partners' names, best first, each once.
- `summary` (text): One line saying what came back.

**Example output**

```json
{
  "count": 3,
  "interactions": [
    {
      "coexpression_score": 0.0,
      "cooccurrence_score": 0.0,
      "database_score": 0.9,
      "directed": false,
      "experimental_score": 0.999,
      "fusion_score": 0.0,
      "neighborhood_score": 0.0,
      "protein_a": "TP53",
      "protein_b": "SIRT1",
      "score": 0.999,
      "sign": null,
      "string_id_a": "9606.ENSP00000269305",
      "string_id_b": "9606.ENSP00000212015",
      "textmining_score": 0.886
    },
    {
      "coexpression_score": 0.052,
      "cooccurrence_score": 0.0,
      "database_score": 0.9,
      "directed": false,
      "experimental_score": 0.999,
      "fusion_score": 0.0,
      "neighborhood_score": 0.0,
      "protein_a": "TP53",
      "protein_b": "MAPK1",
      "score": 0.999,
      "sign": null,
      "string_id_a": "9606.ENSP00000269305",
      "string_id_b": "9606.ENSP00000215832",
      "textmining_score": 0.68
    },
    {
      "coexpression_score": 0.075,
      "cooccurrence_score": 0.0,
      "database_score": 0.75,
      "directed": false,
      "experimental_score": 0.999,
      "fusion_score": 0.0,
      "neighborhood_score": 0.0,
      "protein_a": "TP53",
      "protein_b": "AURKA",
      "score": 0.999,
      "sign": null,
      "string_id_a": "9606.ENSP00000269305",
      "string_id_b": "9606.ENSP00000216911",
      "textmining_score": 0.679
    }
  ],
  "network_type": "functional",
  "partner_count": 3,
  "partners": [
    "SIRT1",
    "MAPK1",
    "AURKA"
  ],
  "summary": "3 functional partners of TP53: SIRT1, MAPK1, AURKA"
}
```

### Get network image

Get links to a picture of the proteins' network (PNG, high resolution PNG or SVG), ready to put in an email, a message or a report, and to the network's page on string-db.org.

**Inputs**

- Proteins (`identifiers`, list, required): The proteins, as gene names (TP53), UniProt accessions (P04637) or STRING IDs (9606.ENSP00000269305), one per line. STRING IDs from 'Map identifiers to STRING' are fastest.
- Organism (`species`, text, required): The proteins' organism. Pick one, or type its NCBI taxon ID (9606 for human) or name. 'Find species' looks up any other.
- Minimum confidence (`required_score`, text): How sure STRING must be of an interaction to include it. Leave empty for STRING's default, usually medium. One of: `900` (Highest confidence (0.900)), `700` (High confidence (0.700)), `400` (Medium confidence (0.400)), `150` (Low confidence (0.150)).
- Network type (`network_type`, text): Functional takes every kind of evidence that proteins work together; physical, only that they bind; regulatory, which one regulates the other. Defaults to functional. One of: `functional` (Functional associations (all evidence)), `physical` (Physical interactions (proteins that bind)), `regulatory` (Regulatory (who regulates whom)). Default: `functional`.
- Extra proteins (`add_nodes`, number): Widen the network with this many of the proteins' closest partners, 0 to 500. With a single protein, STRING adds 10 unless told otherwise. From 0 to 500.
- Line style (`edge_style`, text): How the lines between proteins are drawn. Defaults to coloured by kind of evidence. One of: `evidence` (Coloured by kind of evidence), `confidence` (Thicker for higher confidence), `actions` (Coloured by molecular action). Default: `evidence`.
- Hide unconnected proteins (`hide_disconnected`, true/false): Leave out proteins with no interaction shown.

**Outputs**

- `high_res_image_url` (text): A high-resolution PNG picture of the network.
- `image_url` (text): A PNG picture of the network.
- `page_url` (text): The network's page on string-db.org, to explore it.
- `protein_count` (number): How many proteins were asked about.
- `summary` (text): One line saying what came back.
- `svg_url` (text): The picture as an SVG (vector) file.

**Example output**

```json
{
  "high_res_image_url": "https://version-12-5.string-db.org/api/highres_image/network?identifiers=9606.ENSP00000269305%0D9606.ENSP00000258149%0D9606.ENSP00000384849%0D9606.ENSP00000278616&species=9606&required_score=400&caller_identity=lodol",
  "image_url": "https://version-12-5.string-db.org/api/image/network?identifiers=9606.ENSP00000269305%0D9606.ENSP00000258149%0D9606.ENSP00000384849%0D9606.ENSP00000278616&species=9606&required_score=400&caller_identity=lodol",
  "page_url": "https://version-12-5.string-db.org/cgi/link?to=65FA81FE3499BE7E",
  "protein_count": 4,
  "summary": "Network picture of 4 proteins, ready to open",
  "svg_url": "https://version-12-5.string-db.org/api/svg/network?identifiers=9606.ENSP00000269305%0D9606.ENSP00000258149%0D9606.ENSP00000384849%0D9606.ENSP00000278616&species=9606&required_score=400&caller_identity=lodol"
}
```

### Get similarity scores

Get how similar the proteins of a set are to one another, as Smith-Waterman bit scores (50 or more), to spot paralogs. Each protein is also scored against itself, the most it can score.

**Inputs**

- Proteins (`identifiers`, list, required): The proteins, as gene names (TP53), UniProt accessions (P04637) or STRING IDs (9606.ENSP00000269305), one per line. STRING IDs from 'Map identifiers to STRING' are fastest.
- Organism (`species`, text, required): The proteins' organism. Pick one, or type its NCBI taxon ID (9606 for human) or name. 'Find species' looks up any other.

**Outputs**

- `count` (number): How many scores are listed.
- `scores` (list): Every pair scoring 50 or more, each way round, and each protein against itself.
- `summary` (text): One line saying what came back.

**Example output**

```json
{
  "count": 2,
  "scores": [
    {
      "bitscore": 616.3,
      "protein_a": "CDK2",
      "protein_b": "CDK2",
      "same_protein": true,
      "string_id_a": "9606.ENSP00000266970",
      "string_id_b": "9606.ENSP00000266970",
      "taxon_id_a": "9606",
      "taxon_id_b": "9606"
    },
    {
      "bitscore": 406.8,
      "protein_a": "CDK2",
      "protein_b": "CDK1",
      "same_protein": false,
      "string_id_a": "9606.ENSP00000266970",
      "string_id_b": "9606.ENSP00000378699",
      "taxon_id_a": "9606",
      "taxon_id_b": "9606"
    }
  ],
  "summary": "1 similar pair; most similar: CDK2 and CDK1 (bit score 406.8)"
}
```

### Get version

Get STRING's current release and the release Lodol's STRING steps read, and whether they are the same.

**Inputs**: None.

**Outputs**

- `address_used` (text): The address Lodol calls for it.
- `stable_address` (text): The address that always serves that release.
- `summary` (text): One line saying what came back.
- `up_to_date` (true/false): True when Lodol reads the current release.
- `version` (text): STRING's current release, such as 12.5.
- `version_used` (text): The release Lodol's STRING steps read.

**Example output**

```json
{
  "address_used": "https://version-12-5.string-db.org",
  "stable_address": "https://version-12-5.string-db.org",
  "summary": "STRING 12.5 is current, and Lodol reads it",
  "up_to_date": true,
  "version": "12.5",
  "version_used": "12.5"
}
```

### Map identifiers

Turn gene names, synonyms and UniProt accessions into STRING's protein IDs, with each protein's name and description. Use the IDs in the other STRING steps: they answer faster and mean exactly one protein. Identifiers STRING doesn't know are listed rather than failing the step.

**Inputs**

- Proteins (`identifiers`, list, required): The proteins, as gene names (TP53), UniProt accessions (P04637) or STRING IDs (9606.ENSP00000269305), one per line. STRING IDs from 'Map identifiers to STRING' are fastest.
- Organism (`species`, text, required): The proteins' organism. Pick one, or type its NCBI taxon ID (9606 for human) or name. 'Find species' looks up any other.

**Outputs**

- `all_found` (true/false): True when every identifier was mapped.
- `count` (number): How many identifiers were mapped.
- `matches` (list): The identifiers STRING mapped, in the order given.
- `preferred_names` (list): The proteins' names in STRING, in the same order.
- `string_ids` (list): The STRING IDs alone, in order: pass them to the other STRING steps.
- `summary` (text): One line saying what came back.
- `unmatched` (list): The identifiers STRING could not map.
- `unmatched_count` (number): How many identifiers could not be mapped.

**Example output**

```json
{
  "all_found": false,
  "count": 2,
  "matches": [
    {
      "annotation": "Cellular tumor antigen p53; Acts as a tumor suppressor in many tumor types.",
      "preferred_name": "TP53",
      "query": "TP53",
      "species_name": "Homo sapiens",
      "string_id": "9606.ENSP00000269305",
      "taxon_id": "9606"
    },
    {
      "annotation": "E3 ubiquitin-protein ligase Mdm2; E3 ubiquitin-protein ligase that mediates ubiquitination of p53/TP53, leading to its degradation by the proteasome.",
      "preferred_name": "MDM2",
      "query": "MDM2",
      "species_name": "Homo sapiens",
      "string_id": "9606.ENSP00000258149",
      "taxon_id": "9606"
    }
  ],
  "preferred_names": [
    "TP53",
    "MDM2"
  ],
  "string_ids": [
    "9606.ENSP00000269305",
    "9606.ENSP00000258149"
  ],
  "summary": "Mapped 2 of 3 identifiers; no match for notagene123",
  "unmatched": [
    "notagene123"
  ],
  "unmatched_count": 1
}
```

### Run functional enrichment

Find the functions, pathways and other terms a set of proteins has more of than chance would give: GO processes, functions and components, KEGG, Reactome and WikiPathways, diseases, tissues, domains and papers. Each term comes with its FDR, strength and the proteins that carry it, most significant first. Use two or more proteins: with one, STRING adds its 10 closest partners.

**Inputs**

- Proteins (`identifiers`, list, required): The proteins, as gene names (TP53), UniProt accessions (P04637) or STRING IDs (9606.ENSP00000269305), one per line. STRING IDs from 'Map identifiers to STRING' are fastest.
- Organism (`species`, text, required): The proteins' organism. Pick one, or type its NCBI taxon ID (9606 for human) or name. 'Find species' looks up any other.
- Categories (`categories`, list): Only terms from these categories. Leave empty for all of them. One or more of: `Process` (GO biological process), `Function` (GO molecular function), `Component` (GO cellular component), `KEGG` (KEGG pathways), `RCTM` (Reactome pathways), `WikiPathways`, `Hallmark` (Hallmark gene sets (MSigDB)), `Keyword` (UniProt keywords), `Pfam` (Pfam domains), `InterPro` (InterPro domains and features), and 13 more.
- Maximum FDR (`max_fdr`, text): Only terms with a false discovery rate at or below this. Leave empty for STRING's default. One of: `0.05` (0.05 (5%)), `0.01` (0.01 (1%)), `0.001` (0.001 (0.1%)).
- Minimum proteins per term (`min_gene_count`, number): Only terms carried by at least this many of the proteins. At least 1.
- Background proteins (`background_identifiers`, list): The proteins your set was picked from (every protein your experiment could detect), as STRING IDs. Leave empty for the whole genome.
- Most terms (`limit`, number): The most terms to return, 1 to 1,000, most significant first. Defaults to 100. From 1 to 1000.

**Outputs**

- `count` (number): How many terms are listed.
- `summary` (text): One line saying what came back.
- `terms` (list): The enriched terms, most significant first.
- `total` (number): How many terms were enriched before the limit.

**Example output**

```json
{
  "count": 3,
  "summary": "4 enriched terms for 4 proteins; top: Cellular response to gamma radiation (GO:0071480, FDR 3.45e-08)",
  "terms": [
    {
      "background_gene_count": 27,
      "category": "Process",
      "category_name": "GO biological process",
      "description": "Cellular response to gamma radiation",
      "fdr": 3.45e-08,
      "gene_count": 4,
      "genes": [
        "MDM2",
        "TP53",
        "ATM",
        "CDKN1A"
      ],
      "p_value": 2.8e-12,
      "signal": 5.54,
      "strength": 2.86,
      "string_ids": [
        "9606.ENSP00000258149",
        "9606.ENSP00000269305",
        "9606.ENSP00000278616",
        "9606.ENSP00000384849"
      ],
      "term": "GO:0071480"
    },
    {
      "background_gene_count": 73,
      "category": "KEGG",
      "category_name": "KEGG pathways",
      "description": "Platinum drug resistance",
      "fdr": 3.56e-08,
      "gene_count": 4,
      "genes": [
        "MDM2",
        "TP53",
        "ATM",
        "CDKN1A"
      ],
      "p_value": 1.74e-10,
      "signal": 5.07,
      "strength": 2.43,
      "string_ids": [
        "9606.ENSP00000258149",
        "9606.ENSP00000269305",
        "9606.ENSP00000278616",
        "9606.ENSP00000384849"
      ],
      "term": "hsa01524"
    },
    {
      "background_gene_count": 75,
      "category": "KEGG",
      "category_name": "KEGG pathways",
      "description": "p53 signaling pathway",
      "fdr": 3.56e-08,
      "gene_count": 4,
      "genes": [
        "MDM2",
        "TP53",
        "ATM",
        "CDKN1A"
      ],
      "p_value": 1.94e-10,
      "signal": 5.05,
      "strength": 2.42,
      "string_ids": [
        "9606.ENSP00000258149",
        "9606.ENSP00000269305",
        "9606.ENSP00000278616",
        "9606.ENSP00000384849"
      ],
      "term": "hsa04115"
    }
  ],
  "total": 4
}
```

### Run interaction enrichment

Test whether a set of proteins has more interactions among itself than a random set of the same size would (PPI enrichment): a sign they work together. Returns the counts, the p-value and whether it is significant.

**Inputs**

- Proteins (`identifiers`, list, required): The proteins, as gene names (TP53), UniProt accessions (P04637) or STRING IDs (9606.ENSP00000269305), one per line. STRING IDs from 'Map identifiers to STRING' are fastest.
- Organism (`species`, text, required): The proteins' organism. Pick one, or type its NCBI taxon ID (9606 for human) or name. 'Find species' looks up any other.
- Minimum confidence (`required_score`, text): How sure STRING must be of an interaction to include it. Leave empty for STRING's default, usually medium. One of: `900` (Highest confidence (0.900)), `700` (High confidence (0.700)), `400` (Medium confidence (0.400)), `150` (Low confidence (0.150)).
- Background proteins (`background_identifiers`, list): The proteins your set was picked from (every protein your experiment could detect), as STRING IDs. Leave empty for the whole genome.

**Outputs**

- `average_node_degree` (number): The average number of interactions per protein.
- `clustering_coefficient` (number): The average local clustering coefficient.
- `edge_count` (number): How many interactions they have.
- `expected_edge_count` (number): How many interactions a random set of proteins like these would have.
- `node_count` (number): How many proteins are in the network.
- `p_value` (number): How likely so many interactions are by chance.
- `significant` (true/false): True when the p-value is below 0.05.
- `summary` (text): One line saying what came back.

**Example output**

```json
{
  "average_node_degree": 3.0,
  "clustering_coefficient": 1.0,
  "edge_count": 6,
  "expected_edge_count": 3,
  "node_count": 4,
  "p_value": 0.0569,
  "significant": false,
  "summary": "6 interactions where 3 were expected (p = 0.0569): not significantly more than chance would give"
}
```
