We have developed a number of robust QSAR models to predict physicochemical and DMPK endpoints, along with diagnostic tools to assess their prospective prediction performance.  This is not new.  Many companies have developed similar models.  However, QSAR models do no good if they are not used by synthetic chemists on project teams.  To facilitate usage, we have incorporated direct model access and validation tools into an easy to use modular desktop interface that has been deployed to chemists' desktops.  Modular web-based integration with the help of KNIME back end workflows allow for the easy creation of access ports in a variety of desktop tools.  For example, we have implemented QSAR-based property calculators into PyMol, MOE, Vortex, and StarDrop, and exposed the same calculators to our external collaborators via DEGAS(1).   Examples of the impact routine model usage has had on the acceleration of projects toward clinical candidate molecules will be shown.   (1) Man-Ling Lee, et.al., JCIM, 52(2), 278-84 (2012).

What good are QSAR models if chemists don't use them?

248th National Meeting (2014)

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