Pharmgkb: Obtain A List Of Vip Genes And List All Linked Drugs To Them

Question

[User56]

I am looking for a more advanced version of PharmGKB where I could faster get to solving the problem of obtain a list of VIP genes and list all linked drugs to them.

I know about ability to download individual fileshttps://www.pharmgkb.org/resources/downloads_and_web_services.jsp

Namely I need drugs.tsv, genes.tsv and relationships.tsv.

1. Is there a re-hash of PharmgKB somewhere where I could query all 3 tables right away (similar to people accessing human genome via

   mysql --user=genome --host=genome-mysql.cse.ucsc.edu -A  -D hg19 -e 'SELECT.....''
   

2. Or even some code in R (or other language) which loads it automatically.. something like:

   dataname='drugs'
   temp <- tempfile()
   download.file(paste("http://www.pharmgkb.org/commonFileDownload.action?filename=",dataname,".zip",sep=""),temp)
   drugs <- read.csv(unz(temp, ...)
   

I am thinking that I am not the first one trying to work with their data.

OR any other technology to nail the query in the question title:

E.g.: I saw an OWL download here: http://code.google.com/p/pharmgkb-owl/ but is there anyoen using some OWL query language to achieve this query results: "obtain a list of VIP genes and list all linked drugs to them"

[Fidel]

I don't know any direct solution. What I do is parse the files they provided and feed them into a local SQL database.

[Fidel]

I don't know any direct solution. What I do is parse the files they provide and feed them into a local SQL database.

[User56]

@Fidel: would you mind sharing this strategy. E.g., create statement for all their data files.

Answer 1

[B. Arman Aksoy]

Their OWL dump follows BioPAX specificiations, so if you need to parse those files you can use Paxtools to do run custom queries.

That being said, as far as I know, even OWL files do not contain information about variants and their relative drug associations, that is the relationships.tsv. Again, AFAIK, that file is not re-distributable, so you should expect to find it anywhere else but PharmGKB. You can simply ask for a license to obtain the file and when you have the file, it is not that hard to combine all those tsv's and query drugs for genes of interest.

Probably slightly off-topic, but if you don't need the variant-based information but only gene-centric drug-target assoaciations, you can simply query your genes of interest in a different service like DGIdb, Drugbank or etc.