VarDrugPub :   Variant-Gene-Drug Relation Database

  VarDrugPub is a database that contains genomic variants and pharmacogenomic information. There are numerous research studies published in PubMed about the association/relationship between drug responses and somatic mutations. We use a deep learning-based text mining technique to identify mutation-gene-drug relations in all the PubMed articles. You can use VarDrugPub at

VarDrugPub Database :


  • Extracted Mutataion-Gene-Drug relations with PMID and diseases : DOWNLOAD


  • Total number of PMIDs in the data : 5,685
  • Total number of Sentence(with PMID)-Drug-Mutation-Gene in the data : 11,112
  • Total number of unique Mutation-Drug-(Gene) pairs in the data : 5,712
  • Total number of unique mutations in the data : 3,591
  • Total number of unique genes in the data : 1,077
  • Total number of unique drugs in the data : 901
  • (Last updated on April 25, 2017)


Kyubum Lee, Byounggun Kim, Yonghwa Choi, Sunkyu Kim, Wonho Shin, Sunwon Lee, Sungjoon Park, Seongsoon Kim, Aik Choon Tan, and Jaewoo Kang: Deep learning of mutation-gene-drug relations from the literature. BMC Bioinformatics, 2018. DOI: 10.1186/s12859-018-2029-1 [Full Text]

Contact - for bugs, comments and questions

Kyubum Lee: kyubumlee_at_korea_dot_ac_dot_kr
Byounggun Kim: byounggun_at_korea_dot_ac_dot_kr
Aik-Choon Tan: aikchoon.tan_at_ucdenver_dot_edu
Jaewoo Kang: kangj_at_korea_dot_ac_dot_kr

Last updated on May 23, 2017