Our research program focuses on the development of computational methods for cryo-electron microscopy (cryo-EM). Many of our current developments are taking advantage of current advances in Artificial intelligence and deep learning. Obtaining atomic-resolution macromolecular maps requires deep expertise on cryo-EM. The quality of the data collected depends highly on the expertise of the operator running the microscope. Currently, only few institutions around the world have the required know-how. In addition, appropriate computational tools to analyze challenging cryo-EM samples are still underdeveloped. Then, samples that exhibit high flexibility or multiple conformations are difficult to reconstruct at atomic-resolution. Additionally, current computational methods cannot fully validate resulting cryo-EM structures. Our research is leading the transformation of cryo-EM into a reliable, high-throughput and high-resolution technique that is accessible to the broad scientific community. In this area, We are pursuing the following aims:

Aim 1.1 Developing a robust automated data collection pipeline.

Aim 1.2 Extracting high-resolution information from heterogeneous samples.

Aim 1.3. Developing reliable structure validation methods.

This work will provide computational tools for structural biologists to advance their research programs by allowing them to extract critical information from their cryo-EM data. High-throughput tools will support groups with moderate expertise in cryo-EM to use this technique at its full potential. Novel validation approaches will prevent mistakes and publication of incorrect structures. Finally, innovative methods to analyze the conformational variability will provide new and essential information to infer the macromolecular functions.