We describe algorithms for the rapid and automatic extraction of matched molecular pairs (MMP) from large chemical data sets, and for converting these pairs into structural transformation rules that can be used to drive lead optimization.  These algorithms, implemented in our MedChem StudioTM software application, are part of a broader informatics platform developed to assist chemists with molecule design.  We compare the MMP-based design approach to one using rules based on simple bioisosteric transformations, focusing in particular on predictions of several ADMET properties for a number of literature data sets.  We demonstrate that the effectiveness of the MMP-based approach depends heavily on the amount of local structural information incorporated into each transformation rule, and we discuss a novel approach to allow the amount and nature of this information to be controlled by the software user.  Finally, we present a novel graphical interface called SAR Networks for visualizing and analyzing large numbers of MMP at once.

Improving potency and ADMET properties using matched molecular pair analysis

244th National Meeting (2012)

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