hmdb-database

What it does

Parses HMDB (Human Metabolome Database) local XML to retrieve metabolite information, chemical properties, biological context, disease associations, spectral data, and cross-database mappings. It operates without a REST API and relies on a local XML download (~6 GB). For drugs and live lookups, it directs to drugbank-database-access and pubchem-compound-search respectively.

How it works

• Uses Python's ElementTree to parse hmdb_metabolites.xml and provides utilities to:
  • Build a metabolite lookup index by HMDB ID or name, and retrieve fields like accession, name, chemical_formula, MW, SMILES, InChI, and inchikey.
  • Extract chemical properties (e.g., molecular weight, formula, SMILES, InChI, inchikey, state).
  • Retrieve classification from taxonomy (kingdom, class, etc.).
  • Gather biological context: pathways, biofluid/tissue/cellular locations, and enzymes with protein mappings.
  • Obtain disease associations with OMIM IDs and PubMed references.
  • Access concentration data (normal/abnormal) filtered by biospecimen (e.g., Blood).
  • Extract spectral data: MS and NMR peak lists by spectrum type, with peak details.
  • Compile cross-database mappings to KEGG, PubChem, ChEBI, DrugBank, etc.
  • Provide examples and quick-start snippets for typical tasks.
• The skill includes concrete code patterns for setup, lookup, and each data facet, showing how to parse elements and attributes using the HMDB XML namespace.

When to use it

• When you need metabolite information by HMDB ID or name and require chemical properties (MW, formula, SMILES), biocontext (pathways, enzymes), disease associations, biomarker data, or spectral data from the local HMDB XML.
• When you must map HMDB IDs to KEGG, PubChem, ChEBI, DrugBank, or other databases.
• When live web queries are not available and a local XML workflow is required.

What it can touch

• The skill uses Python code to parse the local XML file hmdb_metabolites.xml and accesses fields via the HMDB namespace (e.g., hmdb:accession, hmdb:name, hmdb:smiles, hmdb:pathways, hmdb:diseases, hmdb:protein_associations, hmdb:spectra, hmdb:kegg_id, etc.).
• It relies on the local file (no REST API) and does not perform network requests itself within the provided routines.

Caveats

• Dependency on a local HMDB XML download (~6 GB uncompressed) and appropriate Python packages (lxml or xml.etree.ElementTree, pandas).
• Not all entries have every field populated (e.g., some spectra, enzyme associations, or concentrations may be missing).
• No explicit handling of licensing beyond noting the CC-BY-4.0 license in the description; usage should comply with HMDB and referenced resources.