OpenFold provides insights into AlphaFold2’s learning behavior
Nature Biotechnology | June 17, 2024
AlphaFold2 has demonstrated impressive utility in predicting protein structure, but the lack of code transparency makes it challenging to understand the model’s learning process or re-train the code for specialized applications. This Research Highlight discusses OpenFold (published in Nature Biotech), a new open-source implementation of AlphaFold2 from LSP fellow Nazim Bouatta and Mohammed AlQuraishi. OpenFold performs with similar accuracy to AlphaFold2 and is more accessible and customizable for the broader protein folding community.
Latest News
October 30, 2025
3D Imaging Maps the Hidden Landscape of Human Tumors
November 4, 2025
Tumor Sight Selected as Biomedical Picture of the Day
October 29, 2025
Paper Talk Podcast – Highly Multiplexed 3D Tumor Profiling
October 15, 2025
Kristin Qian presents work in spatial profiling at European Society for Spatial Biology conference