The use of QSAR technologies is pervasive in most cheminformatics areas. Unfortunately,published QSAR models are rarely provided in an easy to reuse format. Guided by ourexperience with the ChemBench platform (chembench.mml.unc.edu) and utilizing pluginAPI capabilities offered by the NIH Bioassay Research Database (BARD) framework, wehave deployed QSAR model development and distribution option in an open-science, open-access format. We will describe the details of deploying the QSAR modeling capability (usingfreely available CDK descriptors and publicly available machine learning approaches) as aBARD plugin and highlight the challenges in building a predictive model within BARD. Wethen describe a few simple usages of QSAR module as a BARD plugin. The use of BARD formodel deployment in close proximity to the chemogenomic data stored in BARD enables thecommunity to gain access to novel predictive methods in a uniform fashion, leading to enhancedreuse and improved reproducibility.

QSAR modeling on the web

246th National Meeting (2013)

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