Pittsburgh, Pennsylvania, United States
Robert F. Murphy is Professor of Computational Biology Emeritus in the School of Computer Science at Carnegie Mellon University. He was the Ray and Stephanie Lane Professor of Computational Biology and Professor of Biological Sciences, Biomedical Engineering, and Machine Learning at Carnegie Mellon until his retirement in May 2021. He founded the Computational Biology Department in the School of Computer Science at Carnegie Mellon and served as its head from 2009 to 2020. He also cofounded, and served on the board of directors of, Quantitative Medicine, LLC, which was acquired by Predictive Oncology Inc. in July 2020. He has been an editor for the three main bioinformatics journals, a member of the National Advisory General Medical Sciences Council and the NIH Council of Councils, and a member of a number of external advisory boards. He is a Fellow of the IEEE and the American Institute of Medical and Biological Engineering. His interests and experience are in many aspects of machine learning, computational biology and biomedical research, especially in machine learning methods for biomedical image analysis and modeling and AI systems for driving experimental science.
I am Professor of Computational Biology Emeritus in the School of Computer Science at Carnegie Mellon University. I was the Ray and Stephanie Lane Professor of Computational Biology and Professor of Biological Sciences, Biomedical Engineering, and Machine Learning at Carnegie Mellon until my retirement in May 2021. I founded the Computational Biology Department in the School of Computer Science at Carnegie Mellon and served as its head from 2009 to 2020. I also cofounded, and served on the board of directors of, Quantitative Medicine, LLC, which was acquired by Predictive Oncology Inc. in July 2020. I have been an editor for the three main bioinformatics journals, a member of the National Advisory General Medical Sciences Council and the NIH Council of Councils, and a member of a number of external advisory boards. I am a Fellow of the IEEE and the American Institute of Medical and Biological Engineering. My interests and experience are in many aspects of machine learning, computational biology and biomedical research, especially in machine learning methods for biomedical image analysis and modeling and AI systems for driving experimental science. I am open to stimulating new challenges.
The Computational Biology Department within the School of Computer Science was originally named the Lane Center for Computational Biology. It became a department in September 2009 and its name was changed in January 2015.
REDUCING THE DIMENSIONALITY OF BIG SCIENCE Quantitative Medicine will dramatically reduce the time, cost and risk of discovering new therapeutic drugs by predicting: the main effects of drugs on target molecules that mediate disease; the effects of drugs on other molecules or pathways in the body that could mediate adverse effects; as well as the interaction of these with underlying genetic variations. The platform identifies similarities in relationships of drug candidates screened against a diverse matrix of pathogenic, cellular, molecular and/or systems biology targets. By iteratively adding new data from other existing research or additional experiments, the predictive model is improved. More accurate predictions can be made for previously unobserved effects of drug candidates on molecular targets. This scientifically proven approach reduces time sink of the hit-to-lead and lead optimization stages by 1.75 years, improves ROI, and captures up-front cost savings. Functionality includes small molecule and large molecule drug discovery, off-label re-purposing, predicting causes of downstream attrition and adverse effects. Unlike previous efforts which prematurely attempted to replace experimentation with computational predictions, our drug discovery methodology supplements and complements existing methods by cost effectively directing experimentation using an empirical process. Experimentation is optimized for both the hit-to-lead and lead optimization stages. Experiment selection can result in accurate predictions or observations for the most relevant relationships, while only observing a relatively small fraction of them. We accurately predict what we cannot observe. Consequently, the number of experiments needed to know if a drug is a viable candidate to address a selected target can be reduced by at least an order of magnitude. This enables a more complete characterization of different compounds' effects than otherwise affordable.