Mining lexico-syntactic patterns to extract chemical entities with their associated properties

Abstract

Specific information on newly discovered compound is often difficult to be found in chemical databases. The chemical and drug literature is very rich with the information resulted from new chemical synthesis. This paper presents a survey on the types of approaches that have been used to extract information associated with chemical compounds from chemical and drug text. Thereafter, it gives a description for a novel pattern-based extraction method to be developed in the future taking into account specific types of information associated with chemical compounds not explored before in the automated extraction from a text. The paper focuses on the extraction of the properties that influence the bioavailability of drug candidates' compounds. The result of this study can help the database curators in compiling the drug related chemical databases and the researchers to digest the huge amount of textual information which is growing rapidly.