G-protein coupled receptors (GPCRs) form one of the largest gene families of membrane bound proteins. They play a critical role in physiology and in the pathology of many diseases. For this reason, more than 50% of the drugs on the market target GPCRs. However, these drugs have serious side effects. Since GPCRs are membrane bound it is difficult to crystallize them.
In my laboratory, we focus on developing computational methods for predicting the three dimensional structure of GPCRs using minimal experimental data. These computational methods are used to elucidate the differences between the subtypes of GPCRs since the sequence identity of GPCRs prevent the use of homology modeling for this purpose. Further, we predict the binding sites of agonists and antagonists to these receptors using docking algorithms. The predicted binding site and calculated binding energies are then compared with the available experimental data on mutagenesis and ligand binding affinities. These computational predictions greatly aid the experimental screening and discovery of new ligands that could be used either as drugs, or as probes for understanding the physiology of the receptors. We specialize particularly on GPCRs that are cancer targets. I have collaborations with experimental groups and pharmaceutical companies for validating the predictions made using the computational methods.
Another focus of my research is developing hierarchical molecular dynamics algorithms for performing long time scale (microseconds) simulations to study the conformational changes that lead to activation of GPCRs. These constrained dynamics algorithms scale linearly with the number of degrees of freedom and hence would be useful is pushing towards microsecond level simulations. We also apply these long time scale dynamics methods to study the stability of DNA crossover nanostructures. These algorithms are developed in collaboration with scientists at Jet Propulsion Laboratory/Caltech.