Dr. Thomas Maxwell Kaiser is the Molecular Pharmacology Highlighted Trainee Author for the February 2018 issue.
Dr. Kaiser is being recognized for work that he completed as a postdoctoral fellow in the laboratory of Dennis Liotta at Emory University. The Molecular Pharmacology article that earned his selection as a Highlighted Trainee Author is titled “The Bioactive Protein-Ligand Conformation of GluN2C-Selective Positive Allosteric Modulators Bound to the NMDA Receptor.”
Dr. Kaiser’s areas of research are antiviral and anticancer medicinal chemistry, neuropharmacology, and machine learning. He is currently continuing the machine learning work that he began in the Liotta group at Emory, while working toward a medical degree at the University of Oxford. He has devised a technique for machine learning algoriothim generation that can accelerate the discovery of selective therapeutic compounds. He acknowledges the wonderful team in the Liotta group working with him to test the viability of his models.
In his own words, Dr. Kaiser’s sees the potential impact of his current research.
“One of the things that always interested me about chemistry is, historically, chemistry emerges when a physical problem becomes too complex for the mathematical approach of physics. This is especially true in organic synthesis or medicinal chemistry where the scientist is frequently left to assemble empirical evidence to build chemical models about the behavior of molecules (e.g., reactivity trends, structure activity relationships). Often it is chemical intuition in light of chemical models that drives the selection of new compounds for biological testing or the selection of new conditions for a desired transformation. To compute their behavior a priori is frequently either too expensive or not currently possible as the physical models underpinning those calculations are incomplete or nonexistent.
Dr. Kaiser's current work seeks to change that: "Using the massive amount of chemical information in the literature, we can let machine learning build a model for us which we can use to make predictions in a chemical or biological space of interest. Interestingly, machine learning can do this without us having to know the details of the model! We have done this for potency on several biological targets, but, if we can solve the issues of data noise and data set completeness, we could accelerate the generation of quality drug development candidates."
When not in the lab, Dr. Kaiser enjoys cooking, cycling, and a good night at the pub.