Introducing linkedSPLs (linked structured product labels) - bio2rdf

To whom it may concern,


I am Yifan Ning, a system analyst in DBMI, University of Pittsburgh. I hope i can introduce a new project, linkedSPLs, in bio2rdf (github:

https://github.com/bio2rdf/bio2rdf-scripts/tree/release3/linkedSPLs).


In bio2rdf repository, there is a linkedSPLs project that holding all sections of FDA-approved drug package inserts from DailyMed and drug mappings from UMLS, Drugbank, etc. Also, pharmacogenomics and drug-drug interactions annotations are available as well.

We saved all graphs in RDF and provides example queries for join varies graphs to pulls back linked information.

There are seperated five RDF graphs created in linkedSPLs:

(1) linkedSpls core graph
contains all SPLs (structured product labelings) label sections with metadata. They are directly comes from Dailymed product label in XMLs.

(2) Active moiety graph
It's a RDF format dataset reprensting all active ingredients with mappings of RxNORM, ChEBI, Drugbank, DrOn, NDFRT and OMOP.

(3) Pharmgx Data graph
It has pharmacists's annotations for all dailymed label sections

(4) Drug-drug-interactions annotations graph that holding DDIs annotations across dailymed label sections.

(5) Adverse drug reactions
ADR data from SPLs will be provided in the OHDSI Knowledge Base (LAERTES)

Applications for now:
We already uploaded all graphs into public facing virtuoso endpoint for execute varies example queries.

Endpoint at "http://dbmi-icode-01.dbmi.pitt.edu:8080/sparql"

Sample queries are available here "https://github.com/bio2rdf/bio2rdf-scripts/blob/release3/linkedSPLs/example-queries-bio2rdf-linkedSPLs-RDF-01092015..txt"

For these sample quries, please notice that the shorten URL for each queries storing results for query above. (Copy and paste to browser would works)

There is README file in linkedSPLs/LinkedSPLs-update/README that describes how to update those graphs. An Ant project will taking care of getting varies kinds of mappings and stores them in Mysql.

I will be great appreciate if you have any comments and opinions.

Yifan,

DBMI, University of Pittsburgh