Target prediction for unknown compounds or compounds which shows activity in phenotype screening is an important step in understanding Mode of action and Off-Targets. Machine learning methods have proved lot of assurance in fishing out potential targets on the basis of chemical structure. The present work we built Insilco Target specific models of more than 2200 targets using laplacian modified bayesian classifier and assess its performance based on different datasets. We integrated different databases and extracted activity information from GVK, CHEMBL21 , Thomson Reuters , WOMBAT, AUREUS and in-house datasets. We classified the targets onto different classes based on the panther classification. Finally, we assess the performance of the models based BEDROC and AUCPR. The target prediction tool is used by the HTS group in recognition of targets for phenotype-based screenings.

AbbVie target prediction tool

253rd National Meeting (2017)

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