SUPR
Data Driven Methods for Cryo-EM Reconstruction Based on Topological Neural Networks
Dnr:

NAISS 2025/22-625

Type:

NAISS Small Compute

Principal Investigator:

Jonathan Krook

Affiliation:

Kungliga Tekniska högskolan

Start Date:

2025-06-03

End Date:

2026-07-01

Primary Classification:

10105: Computational Mathematics

Webpage:

Allocation

Abstract

Cryo-EM is a method for 3D imaging of biological macromolecules through transmission electron microscopes. Although data driven methods for solving the Cryo-EM reconstruction problem already exist, none have been able to provide an entirely satisfying solution. We want to test data driven methods for 3D reconstruction that leverage the geometry of the object being studied. Specifically we will use neural networks from a broad class of models referred to as topological neural networks.