First-visit guidance

Documentation and help

This page consolidates the quick start, search guidance, data model, curation methods, identifier policy and frequently asked questions.

Quick start

1. Enter a query

Use a gene symbol, alias, TDB accession, BioGRID interaction ID or PubMed identifier.

2. Review matches

Filter by species, relation type, directionality, curation origin and evidence level.

3. Verify evidence

Open a TDB record, inspect qualification text and follow every supporting source identifier.

Search guide

Keep the first search simple, then expose advanced filters after results load. Suggested input types:

Input	Meaning	Expected behavior
`GENE1`	Official symbol or alias	Return every curated combination containing the matched gene.
`TDB000001`	Stable TripleDB accession	Open the exact curated record.
`BIOGRID:1113787`	Source interaction ID	Show the source record and all linked TDB accessions.
`PMID:12345678`	Publication source	Return curated records and source evidence from that paper.

Usability rule:

Display examples next to the search box and accept identifiers with or without common prefixes where ambiguity is low.

Data model

A TripleDB record represents one curated higher-order genetic relationship. It may aggregate several BioGRID source records, include manually curated literature-only evidence, or combine both.

ENTITY

TDB record

Stable identity and record-level version.

BIOLOGY

Relationship

Genes, relation type, direction, phenotype and condition.

EVIDENCE

Sources

BioGRID IDs, PMID/DOI, qualification and source location.

CURATION

History

Origin, confidence, notes, revisions and status.

Directionality and gene roles

Symmetric

Set-based interaction

Use for relationships such as synthetic lethality when the source does not define a causal ordering among genes.

Directional

Background → modifier

Use when a gene or perturbation rescues, suppresses or enhances a defined background phenotype.

Ambiguous

Roles unresolved

Retain the relationship and evidence text while explicitly avoiding an unsupported interpretation.

Provenance and evidence linking

Three identifiers must remain conceptually separate:

TDB accession: The stable identifier for the curated high-order interaction record.
BioGRID interaction ID: An external source-record identifier retained as provenance; several may support one TDB record.
PMID / DOI: The publication evidence source. PubMed identifiers may be detected from BioGRID source records or embedded in qualification text for manually curated records.

For download convenience, multiple source IDs may be pipe-delimited in the flat table, while the internal database should normalize them as one source-link row per identifier.

Curation workflow

Freeze and document a BioGRID source release.
Apply reproducible candidate filters to genetic, same-species, triple-related records.
Review all relevant source rows and the publication context.
Aggregate records that support the same higher-order biological conclusion.
Assign relation type, directionality, roles, phenotype, condition and confidence only where supported.
Create a stable TDB accession and preserve all source links.
Run quality control and publish the record within a versioned release.

Do not silently overwrite:

Corrections should preserve the accession and create a visible change-history event. Invalidated records should be deprecated, not recycled.

Identifier and version policy

Every curated high-order record receives one TDB accession.
TDB accessions are never reassigned to a different relationship.
New records append after the highest existing accession.
Minor annotation corrections retain the same accession and increment record history.
Major reinterpretation is documented; splitting or merging records preserves redirects and predecessor links.
Deprecated accessions remain resolvable with a reason and replacement when applicable.

See the proposed release and deprecation policy →

Core field dictionary

TripleDB_ID

Stable accession for one curated TripleDB higher-order interaction record.

Source_ID

BioGRID interaction ID(s) separated by |, or a TripleDB manual source identifier when the record is curated from non-BioGRID literature.

Organism

Organism label used by TripleDB and the alias/source indexes.

Experimental System

Curated interaction phenotype or relation class, such as synthetic lethality, synthetic rescue, phenotypic enhancement or negative genetic interaction.

Gene_A–Gene_G

Ordered gene members in the curated triple or higher-order interaction record. Empty cells indicate that no additional gene member is used for that record.

Qualifications

Preserved evidence text, curator note, source context and embedded literature identifiers such as PUBMED:12345678.

_derived.source_ids

Parsed source identifiers generated for website search and evidence linking.

_derived.genes_indexed

Normalized gene list generated from Gene_A–Gene_G and used by the website gene search.

_derived.interaction_order_class

Interaction-order class used by the interface, including triple, quadruple_or_higher and contextual pair-level records.

_derived.pubmed_ids

PubMed identifiers detected from qualification text when available.

Frequently asked questions

Why does every curated combination need a TDB accession?

Because the aggregated high-order conclusion is a TripleDB object, not any single underlying BioGRID pairwise record. A stable accession allows citation, revision tracking and permanent linking.

Can a record list many BioGRID interaction IDs?

Yes. The website can display them as individual links; a flat-file export may use a documented pipe delimiter. Internally, use a normalized source-link table.

How are PubMed IDs handled?

PubMed identifiers are detected from BioGRID source records and from qualification text in manually curated records. The website uses these identifiers for publication statistics and PubMed links.

Does the first release need an API or network visualization?

No. A clear search, stable record pages, evidence provenance, complete downloads, help and version information are higher priority. An API and advanced visualization can follow after the data model is stable.