How Many Human Drug Targets Are There?

Question

[Andrew Su]

This 2006 paper estimates the number of human targets for approved drugs at 266 (see Table 1). Anyone know of a more recent paper with an updated count?

Answer 1

[Egon Willighagen]

You can also use a SPARQL end point of ChEMBL (version 02) and search for all listed targets. There seems to be a limit of three URLs, and the URLs do not work out well, so here's how to reproduce it... Got to the SPARQL end point linked earlier, and copy/paste the SPARQL bits (excluding the PREFIX bits, just like the third link).

To list all human targets, use this SPARQL:

SELECT DISTINCT ?instance ?organism
WHERE {
  ?instance a <http://pele.farmbio.uu.se/chembl/onto/#Target> ;
    <http://pele.farmbio.uu.se/chembl/onto/#organism> ?organism .
  FILTER regex(?organism, "Homo sapiens")
}

To count the human targets, use this SPARQL:

SELECT count(?instance)
WHERE {
  ?instance a <http://pele.farmbio.uu.se/chembl/onto/#Target> ;
    <http://pele.farmbio.uu.se/chembl/onto/#organism> ?organism .
  FILTER regex(?organism, "Homo sapiens")
}

With this last query, we see that ChEMBL lists 2604 human targets.

Now, the next step is to filter only those targets with approved drugs... I am now checking if this information is available from ChEMBL and will get back on this shortly...

Update: ChEMBL 10 has a list of FDA approved drugs, making the full query possible. I have not RDF-ized version 10 yet, but ping me in a few weeks.

[Pierre Lindenbaum]

cool answer :-)

[Egon Willighagen]

Oh, just wait until I get that federated SPARQL going with some public service... see http://www.semanticoverflow.com/questions/1830/federated-sparql-on-sparlql-org

[Andrew Su]

Agreed, cool answer. Anxiously awaiting the conclusion...

[Egon Willighagen]

Yeah, I had the wrong URIs in my ChEMBL data for Bio2RDF content... fixing that today, and then I can link it up with DrugBank data (and more)...

[Hranjeev]

Doesn't work at the moment?

[Egon Willighagen]

Please try now.

[Hranjeev]

No luck. Maybe I got something wrong?

[Hranjeev]

The http://pele.farmbio.uu.se/chembl/onto/#Target link returns a 404

[Egon Willighagen]

Correct. It's a undereferencable URI. I have not formalized an ontology yet.

[Egon Willighagen]

But you don't need to dereference #Target to use the SPARQL end point.

[Egon Willighagen]

Added links now for the two queries. BTW, the problem was that I updated the RDF store, changing the URIs of the #Target class and #organism predicate. Sorry about that :(

Answer 2

[Mary]

You might find some of the stuff you need at DrugBank: http://www.drugbank.ca/

Possibly also can pull it out of Comparative Toxicology Database with some queries: http://ctd.mdibl.org/

They may both have relevant publications with that kind of stat too.

Answer 3

[Nir London]

You can also check out the "Therapeutic Target Database" in the papar abstract from 2002 they claim it contains entries for 433 targets covering 125 disease conditions along with 809 drugs/ligands directed at each of these targets. Or the "Open access therapeutic target database".

Answer 4

[Fred Fleche]

[?]

http://www.bioinformatics.fr/biology.php?subsection=Drug

Answer 5

[Sophia]

I found that KEGG Drug is a nicely organized resource.

For each drug, you can find entries on target genes (TARGET), genes involved with the metabolism (METABOLISM) of the drug and genes interacting with the drug (INTERACTION) in another way. See an example entry here .