AlphaSAXS: Reconstructing Protein Structure with Physiologically Relevant Conformations from Small Angle X-ray Scattering Data

Feng Yu, Stephanie Prince, Andrew Tritt, Kanupriya Pande, Greg L. Hura, Oliver Ruebel, Susan E. Tsutakawa

Track: Biology: datasets and/or experimental results

Nature Biotechnology: Yes

Keywords: Protein design, high throughput data, small-angle X-ray scattering, experimental data, conformational diversity

TL;DR: AlphaSAXS addresses AlphaFold's limitations in predicting flexible protein structures by incorporating X-ray scattering data.

Abstract: AlphaFold has revolutionized structural biology with accurate protein predictions, yet challenges remain due to the dynamic nature of proteins. Over 40% of human proteins have flexible regions crucial in diseases like Alzheimer's and COVID-19. To overcome AlphaFold's limitations, we integrated small-angle X-ray scattering (SAXS) data, leveraging its ability to provide structural insights on flexible macromolecules. Using computationally generated SAXS data to inform network inputs during fine-tuning and inference, we enhanced AlphaFold to improve conformation predictions for flexible protein regions. This approach can advance our understanding of experimentally guided structural prediction and provides a potential solution for improving computational prediction of physiologically relevant conformations.

Anonymization: This submission has been anonymized for double-blind review via the removal of identifying information such as names, affiliations, and identifying URLs.

Presenter: ~Stephanie_Prince1

Format: Yes, the presenting author will attend in person if this work is accepted to the workshop.

Funding: No, the presenting author of this submission does *not* fall under ICLR’s funding aims, or has sufficient alternate funding.

Submission Number: 101