Nat Comput SciJClub
HDOCK-Multimer: improved structure prediction of large protein complexes by integrating docking and combinatorial assembly.
Nature Computational Science · · Journal Article · Open access
Yao, Ya + more
Abstract ↗AI summary
The abstract is read at the publisher; the summary is JClub's.
HDOCK-Multimer (HDM) significantly outperforms existing methods in predicting large protein complex structures across multiple benchmarks.
- Why it matters: Accurate structure prediction of large protein assemblies is essential for understanding biological functions but remains difficult for current end-to-end approaches like AlphaFold due to size and complex interactions.
- What they did: The authors developed HDM, a framework combining docking and combinatorial assembly, tested on 3 benchmarks with 35 heteromeric complexes, 172 large complexes, and 7 CASP15 targets.
- The result: HDM surpasses state-of-the-art methods, including CombFold, MoLPC, AlphaFold-Multimer, and AlphaFold3, and can predict complex stoichiometry without prior input, enabling improved large protein complex modeling.
The findingWhy it mattersWhat they didThe result
- Open access