Zusammen

Understanding drug metabolism pathways plays an essential role in drug discovery and development. Today's stop on our continuing survey of free chemistry databases and Web services focuses on this problem. The Human Metabolome Database (HMDB) contains information on small organic molecules, their physical properties, and the enzymes that metabolize them.

From the about page:

The Human Metabolome Database (HMDB) is a freely available electronic database containing detailed information about small molecule metabolites found in the human body. It is intended to be used for applications in metabolomics, clinical chemistry, biomarker discovery and general education. The database is designed to contain or link three kinds of data: 1) chemical data, 2) clinical data, and 3) molecular biology/biochemistry data. The database (version 2.0) contains over 6500 metabolite entries including both water-soluble and lipid soluble metabolites as well as metabolites that would be regarded as either abundant (> 1 uM) or relatively rare (< 1 nM). Additionally, approximately 1500 protein (and DNA) sequences are linked to these metabolite entries. ...

HMDB is organized around small molecule drugs and their human metabolites. A large amount of information is available in each entry, including: chemical structure; placement in a chemical taxonomy; links to external databases and preparation; physical properties such as melting point, spectral data in graphical format including 1H, 13H and mass spectra; and predicted spectral properties. Currently, spectral data are available in raster image format only. In addition to physical properties, each entry lists the enzymes involved in a metabolic pathway.

The entry for 2-Furoylglycine gives a good idea of what's available.

HBDB supports many ways to search, including exact structure search, substructure search, plain text search, and FASTA search.

In addition to these searches, HMDB also offers spectral data peak search. For example, this page enables peak seach of HMDB's NMR spectra. Although the form is quite complex, a number of pre-configured examples are available that return illustrative hitsets.

The entire HMDB database, including structures and data, can be downloaded, offering many interesting possibilities for integrating HMDB with other online resources.

Although overall a good concept and implementation, HMDB does have some areas that could be improved. For example, although enzyme data are included along with small molecule entries, these data don't appear to be independently-addressable. This might be useful for finding all small molecules involved with a given protein, for example. Additionally, DMDB doesn't appear to expose the relationship between small molecule and enzyme (e.g., metabolite or substrate). Another problem area was broken links. For example, the 2-Furoylglycine entry has an internal link to "METLIN ID" that didn't work when I tested it.

An important step in designing drugs with better metabolic profiles is to understand the ways existing drugs are metabolized. HMDB offers a solid approach the problem of collecting and making the existing corpus of data available. For those involved in metabolite identification, drug development, or modelling metabolism processes, HBDB is worth looking at.

Kudos

• Free access to all data.
• Links to primary literature.
• Spectral data search.
• Primary spectral data available.

Ideas for Improvement

• Make enzyme records individually-addressable resources as small molecules are. Do the same for chemical taxononomy.
• Expose relationship (metabolite, substrate) between enzyme and small molecule.
• Offer spectral data in machine-readable format such as JCAMP-DX.
• Fix broken links.