PK Corpus
We have annotated abstracts from National Library of Medicine's MEDLINE database. Version 0 data comprises of 1213 articles in total. We divide those articles into two grouops according to the relevance of Drug-Drug interaction (DDI). There are 602 DDI-relevant abstracst and 611 DDI-irrelevant abstracts.


In Version 1.0 data, the pharmacokinetic corpus currently consists of 541 abstracts, these abtstracts were annotated and labelled on the basis of the Pharmacokinetic (PK) relevant MESH terms like (Drug Name,Enzyme name, PK parameters Eg-Clearance).



Figure shows the proportion of drugs metabolized by different CYP families.