PGxCorpus: a Manually Annotated Corpus for Pharmacogenomics

Abstract

Pharmacogenomics (PGx) studies how individual gene variations impact drug response phenotypes, which makes knowledge related to PGx a key component towards precision medicine. A significant part of the state-of-the-art knowledge in PGx is accumulated in scientific publications, where it is hardly usable to humans or software. Natural language processing techniques have been developed and are indeed employed for guiding experts curating this amount of knowledge. But, existing works are limited by the absence of high quality annotated corpora focusing on the domain. This absence restricts in particular the use of supervised machine learning approaches. This article introduces PGxCorpus, a manually annotated corpus, designed for the automatic extraction of PGx relationships from text. It comprises 945 sentences from 911 PubMed abstracts, annotated with PGx entities of interest (mainly genes variations, gene, drugs and phenotypes), and relationships between those. We present in this article the method used to annotate consistently texts, and a baseline experiment that illustrates how this resource may be leveraged to synthesize and summarize PGx knowledge.