The Kavraki Lab has developed a computational tool, known as DINC, that can be used to predict the conformation of a given protein and ligand pair after they bind. This prediction task is non-trivial and becomes even more challenging as the size and complexity of the ligand increase. Predicting the binding modes of larger ligands has important applications in drug development. Much of the research in drug design, including the “designer drugs” that are increasingly used in immunotherapy treatments. Docking tools such as AutoDock and Vina exist that are able to predict the binding modes of smaller molecules with good accuracy and consistency, however the results break down once larger ligands are considered.

In an attempt to improve the accuracy of the final conformation prediction, the DINC tool uses an incremental approach that docks a single fragment of the ligand at a time. At each step of the incremental process a different docking tool is used to actually dock the individual fragment. DINC had been using AutoDock4 for this task, but the lab realized that a different tool, Vina, may help DINC yield better results. Part of my work this summer will be focused on bench-marking the results of DINC using Vina.

Another aspect of my project for the summer is adding some functionality to the DINC program. Some of the docking jobs take a very long time, even when run on a supercomputer, because it is a computationally intense process. I will add functionality to the program to allow a docking job to restart from a mid-point. This will allow a docking job that has been interrupted for whatever reason to be continued without discarding and wasting previous computations.

The final piece of my project will entail exploring how some of the intermediate results produced during a docking job might be analyzed to better understand the docking process and final result. Many files are generated in the process of the docking job that contain information regarding the intermediate stages of docking. The lab is not yet using that information.

You can download my final report on the project here:
Final Report