Deep Learning-Accelerated Crystallography Pipeline

Our project aims to advance the determination of small molecule structures using crystallography, a technique vital for developing new drugs and materials. This process can be complex and time-consuming, especially for challenging cases.

We will use advanced machine learning (ML) techniques to create a user-friendly pipeline, making crystallography faster and more accessible. By training ML algorithms on simulated data, we can improve the accuracy and efficiency of structure determination, even for difficult samples like those with low resolution or disorder.

Our team, including experts from the University of Copenhagen, Durham University, and the MAX IV synchrotron, will develop and test these tools. The goal is to integrate these ML algorithms into open-source software, making crystallography accessible to non-specialists and accelerating scientific discovery.