Uncommon Data Sources for QSAR Modeling
Alexander Tropsha · 2018
This chapter comments on the unusual sources of data and opportunities to develop models from uncommon sources as they highlight the growing diversity of data sources and correspondingly and growing appeal of quantitative structure-activity relationship (QSAR) to such diverse research fields as text mining and medical informatics. One of the first instances of QSAR model development using uncommon input data is provided by the study of chemical hepatotoxicity in collaboration with BioWisdom Ltd. This study was the first application of QSAR modeling and other cheminformatics techniques to observational data generated by the means of automated text mining with limited manual curation, opening up new opportunities for generating and modeling chemical toxicology data. Very recently, this original investigation follows into the use of uncommon electronic sources of input data for QSAR modeling in another study that looked into a specific adverse drug effect known as Stevens-Johnson Syndrome (SJS).