Patents are a widely recognised valuable source of early structure-activity relationship (SAR) information; however, the manual efforts involved in extracting SAR data from patents is costly and exacerbated by the purposely obfuscated nature of patents.  Technological advances in text mining of patents using natural language processing (NLP) approaches has recently resulted in the creation of efficient workflows for the extraction of SAR data. The workflow combines and integrates software capable of converting chemical text into structural representation with NLP software that recognises the wide variety of different ways inhibition data can be described within the text of patents including biological assay data found inside tables. Linguamatics describes use of the I2E software in establishing a triage method that allows medicinal chemists to quickly assess the potential of a patent to be text mined for SAR data.

Automating chemical structure and inhibition data extraction from patents: Text-mining approach

2019 Fall National Meeting

The official archive and repository for all ACS Meetings & Expositions abstracts and select posters or presentations.