Release files
Attached to every release by the release workflows. Sizes are read from the latest release when this page loads.
| File | What it is | Format | Size |
|---|---|---|---|
dismech_nodes.jsonl |
Knowledge-graph nodes: diseases, phenotypes, genes, cell types, processes and the other entities the KB names. | KGX JSON Lines (Biolink Model) | — |
dismech_edges.jsonl |
Knowledge-graph edges between those nodes. | KGX JSON Lines (Biolink Model) | — |
dismech_sepio.jsonl |
Evidence and provenance for the edges: the cited references and quoted snippets behind each one. | JSON Lines (SEPIO) | — |
dismech_nodes.tsv, dismech_edges.tsv |
The same nodes and edges as the two files above, as tab-separated tables with every property kept. Multiple values in one cell are separated by |. Some quoted evidence contains a | of its own (rows quoted from Orphanet and ClinGen tables); there it is written as ¦, so it does not split the quote. Line breaks in quotes become spaces. The JSON Lines files keep the text exactly. |
KGX TSV (Biolink Model) | —, — |
dismech_kgx_metadata.json |
The Biolink Model version the knowledge graph was built against, and its node and edge counts by category and predicate. | JSON | — |
mondo_emc.tsv |
Disease records keyed to MONDO, for Mondo's external-content import. | TSV | — |
kb.zip |
The curated entries themselves: every YAML file under kb/ (disorders, modules, module collections, groupings, trajectories, hypothesis assessments and the rest), the curated gene summaries, and the pathograph node-class vocabulary. It leaves out the copied HGNC and ClinGen gene tables and the hypothesis providers' raw reports and figures. |
Zip of YAML and Markdown | — |
pathographs.zip |
The causal graph of every disorder that has one, as the site draws it, plus disease and gene indexes. Disorders without a MONDO term or without any causal edge have no file. | Zip of JSON | — |
pathographs_cx2.zip |
The same causal graphs as networks for NDEx and Cytoscape, one per disorder with at least one causal edge. | Zip of CX2 JSON | — |
dismech_models_omex.zip |
The SBML models behind the runnable-model pages: one COMBINE archive per model, holding the model and a SED-ML simulation of each scenario, which any SED-ML tool can run. Models written in this repository with their own runner (the Boolean-network and agent-based models), and coupled co-simulations that SED-ML cannot express, are not included. | Zip of COMBINE .omex |
— |
release_manifest.json |
The number of entries in each part of the KB (disorders, modules, groupings and so on) at the release, and the size and SHA-256 checksum of each archive above. | JSON | — |
Experimental release files
Attached to every release, but their contents and identifiers may change without notice.
| File | What it is | Format | Size |
|---|---|---|---|
dismech_maximal_nodes.jsonl, dismech_maximal_edges.jsonl |
The whole KB as one graph: disorders, modules, groupings and trajectories, with every mechanism node of every entry as its own node, plus conformance, comorbidity and grouping-membership edges. Unlike the Biolink files, most predicates and categories are DisMech's own. A mechanism node's identifier is built from its entry's file name and its own name, so renaming either changes it (see #12368). | KGX-style JSON Lines | —, — |
dismech_maximal_metadata.json |
Node and edge counts of the graph above, by category and predicate. | JSON | — |
The knowledge base itself
The curated entries are YAML files, one per disorder, module, grouping and so on, validated against a LinkML schema.
- Browse
kb/on GitHub —kb/disorders/,kb/modules/,kb/groupings/and the rest, at the currentmain. kb.zipfrom the latest release (see the table above) — the curated content only.- Source archive of the latest release (zip) — the whole repository at the release tag, code and caches included, so it is large.
- The LinkML schema (
src/dismech/schema/dismech.yaml) that the YAML follows, also browsable as documentation. - monarch-initiative/dismech-owl — a separate repository for an OWL version of DisMech.
Not yet published
- Pathograph networks on NDEx. The CX2 networks are in
pathographs_cx2.zipabove, but none has been uploaded to NDEx yet. - Embeddings. The vectors behind the embedding maps are computed with a paid API and are not attached to releases.
- HPO-annotation export. A phenotype-annotation file in the HPO
.hpoaformat. It is not attached to releases; build it from a checkout withjust export-hpoa, which writes tooutput/hpoa/.
License and citation
The pages on this site are marked CC-BY 4.0. A license statement specific to the bulk data, and a recommended citation, are not settled yet. Until they are, cite the release you used by its tag and URL, so others can retrieve the same files.
Some records contain material owned by third parties, used under their own terms; LOINC content in particular. See NOTICE.