Simulations Plus, Inc. (NASDAQ: SLP), a leading provider of simulation and modeling software to the pharmaceutical industry, announces the release of ADMET Predictor™ 4.0, a major update to its top-ranked software for the prediction of properties of new molecules from just their structure.
Dr. Robert Fraczkiewicz, product manager for ADMET Predictor and team leader for the ADMET Cheminformatics Team at Simulations Plus, said: “This new version of ADMET Predictor includes extensive improvements to its already-top-ranked models for prediction accuracy and it adds a number of brand new functionalities. These include a total of 15 new predicted properties: new models for metabolism by seven different UGT enzymes, new models for five different liver toxicities, a new model (‘logD’) for how molecules partition between water and a fatty liquid based on measurements for more than 10,000 molecules, a new model for predicting the ratio of the concentration of a new molecule in whole blood compared to only in plasma, and a new model that predicts the likelihood of a compound remaining in a supersaturated condition if it moves from a region of high solubility at one pH to one of lower solubility at a different pH in the gastrointestinal tract. In addition to these 15 new models, we have retrained all of our existing models using the newest molecular and atomic descriptors we generated during our NIH SBIR Phase I grant and additional work we accomplished between the Phase I and our recent Phase II grant award, resulting in improving all of those models. As an enhanced user convenience, we have provided a powerful graphics engine that enables users to ‘see’ in up to 8 dimensions for more insightful visual screening of molecules based on their various properties and molecular characteristics.”
Dr. Fraczkiewicz continued: “In addition to the new and upgraded models, we have also added new functionality to both ADMET Predictor and its embedded model-building program, ADMET Modeler™. In ADMET Predictor, we have developed, under a funded collaboration with Pfizer, the ability for the program to estimate the dose amount that would be required to deliver an effective plasma concentration for a series of new molecules. This capability enables chemists to account for the complex interactions among many properties that affect plasma concentration levels, and hence affect the dose amount required to achieve the desired levels. We have also added an upgraded mini-GastroPlus™ simulation capability that is used for this dose estimation and for calculation of fraction of dose absorbed at fixed dose levels. In addition, a very novel concept of the Structure-Activity Landscape Index (SALI), published recently by top researchers in the field: Drs. Rajarshi Guha and John Van Drie, has been implemented to evaluate a predictive model’s ability to capture ‘activity cliffs’ – a tricky problem wherein very similar molecules have very different properties. To our knowledge, we are the first commercial software vendor to do so.”
Walt Woltosz, chairman and chief executive officer of Simulations Plus, added: “Prior to this release, ADMET Predictor enjoyed the position of being top-ranked for predictive accuracy in every published independent comparison study of the 15-20 competitive programs that are offered today. Interestingly, no competitor has even been consistently in the top four across these studies. That says a lot. With this new release, we not only have added new predictions, but we have also improved our earlier models. We believe our consistent top-ranked performance comes from our unique library of molecular and atomic level descriptors, from our care in making sure the training data are free from errors, and from the sophisticated modeling technologies in ADMET Modeler we’ve developed over the past 7 years. We expect continued growth of ADMET Predictor licenses in this competitive environment as it continues to lead the pack in predictive accuracy, speed, and user convenience.”