Fast and comprehensive structure identification in cryo-ET using deep learning and tomogram optimisation

ongoing

Cryo-electron tomography (cryo-ET) revolutionises cellular structure visualisation, but extracting biological insights is often hindered by limitations of subtomogram averaging (STA). With AlphaFold2's recent breakthroughs in structure prediction, identifying known structures in their 3D cellular context is often more important than structure determination by STA. The TomoFind project aims to significantly improve our ability to identify biomolecular structures within tomographic reconstructions. This goal will be pursued through three interconnected approaches: 1) Developing deep learning tools for rapid and comprehensive identification of known structures in tomograms; 2) Refining data acquisition parameters via mathematical modelling and computational optimisation; and 3) Enhancing reconstruction algorithms to fully leverage advancements from 1) and 2). This highly interdisciplinary project combines applied mathematics, machine learning, and structural biology, and close collaboration between computational and experimental groups ensures practical applicability and maximises chances of success. TomoFind will advance cryo-ET analysis capabilities, enabling investigation of cellular structures and processes at unprecedented resolution and scale. The resulting open-source software will lead to high impact across structural and cell biology research communities. The host institution's world-class technical expertise and state-of-the-art facilities provide an ideal environment for this project. This, combined with the applicant's interdisciplinary background spanning applied mathematics, data science, and computational imaging, forms a strong foundation, crucial for the success of TomoFind. The project will enhance the applicant's skills and network, boosting future prospects for prestigious European research positions. The host institute will also benefit from the project's outcomes, advancing its research capabilities in cryo-ET and structural biology.