- Indico style
- Indico style - inline minutes
- Indico style - numbered
- Indico style - numbered + minutes
- Indico Weeks View
Speaker: Krzysztof Fidelis (UC Davis / CASP)
Topic: From Proteins to Virtual Cells: Advances and Challenges in Computational Structure Prediction – A CASP perspective
Abstract: Recent breakthroughs in protein structure modeling, notably through deep learning methods exemplified by AlphaFold, have profoundly transformed computational structural biology. CASP (Critical Assessment of Structure Prediction) has systematically evaluated these advances, most recently in CASP16 (2024), showcasing significant improvements in predicting protein structures with accuracies comparable to experimental data. I will provide a concise historical overview of key developments in CASP and highlight current results from CASP16. I will also outline the emerging challenges: predicting RNA structures, assessing computational models when clear experimental benchmarks (ground truth) are unavailable, and evaluating integrative modeling methods that rely on diverse experimental data types. I will also explore and invite feedback on extending the CASP assessment approach to foundational models within more complex systems, such as virtual cells or artificial organisms, underscoring critical questions about method assessment in these broader computational biology contexts