Reproduce “Predicting protein structures at any folding stage from sequence” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Lisheng Zhang, Qingsong Li, Zinan Xie, Yanting Yang, Qikui Yang, Peng Lu, Zhican Fu, Wentao Cui and 2 more
FoldScape generates ensembles of protein structures from sequence, conditioned on folding stage, compactness, and secondary-structure inputs. A flow-matching decoder produces Cα traces without protein-specific molecular dynamics simulation. Across 20 human proteins, median Q coverage was 0.722 versus 0.05 for seed-varied AlphaFold 3. This measures breadth, not accuracy or physical probability; coverage and reconstruction were tested on different targets.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.