loading page

Integrative modeling in the age of machine learning: a summary of HADDOCK strategies in CAPRI rounds 47-55
  • +14
  • Victor Reys,
  • Marco Giulini,
  • Vlad Cojocaru,
  • Anna Engel,
  • Xiaotong Xu,
  • Jorge Roel,
  • Cunliang Geng,
  • Francesco Ambrosetti,
  • Brian Jimenes-Garcia,
  • Zuzana Jandova,
  • Panagiotis I. Koukos,
  • Charlotte van Noort,
  • João Teixeira,
  • Siri C. van Keulen,
  • Manon Reau,
  • Rodrigo Vargas Honorato,
  • Alexandre M.J.J. Bonvin
Victor Reys
Universiteit Utrecht Departement Scheikunde
Author Profile
Marco Giulini
Universiteit Utrecht Departement Scheikunde
Author Profile
Vlad Cojocaru
Universiteit Utrecht Departement Scheikunde
Author Profile
Anna Engel
Universiteit Utrecht Departement Scheikunde
Author Profile
Xiaotong Xu
Universiteit Utrecht Departement Scheikunde
Author Profile
Jorge Roel
Universiteit Utrecht Departement Scheikunde
Author Profile
Cunliang Geng
Universiteit Utrecht Departement Scheikunde
Author Profile
Francesco Ambrosetti
Universiteit Utrecht Departement Scheikunde
Author Profile
Brian Jimenes-Garcia
Universiteit Utrecht Departement Scheikunde
Author Profile
Zuzana Jandova
Universiteit Utrecht Departement Scheikunde
Author Profile
Panagiotis I. Koukos
Universiteit Utrecht Departement Scheikunde
Author Profile
Charlotte van Noort
Universiteit Utrecht Departement Scheikunde
Author Profile
João Teixeira
Universiteit Utrecht Departement Scheikunde
Author Profile
Siri C. van Keulen
Universiteit Utrecht Departement Scheikunde
Author Profile
Manon Reau
Universiteit Utrecht Departement Scheikunde
Author Profile
Rodrigo Vargas Honorato
Universiteit Utrecht Departement Scheikunde
Author Profile
Alexandre M.J.J. Bonvin
Universiteit Utrecht Departement Scheikunde

Corresponding Author:[email protected]

Author Profile

Abstract

The HADDOCK team participated in CAPRI rounds 47-55 as both server, manual predictor, and scorers. Throughout these CAPRI rounds, we used a plethora of computational strategies to predict the structure of protein complexes. Of the 10 targets comprising 24 interfaces, we achieved acceptable or better models for 3 targets in the human category and 1 in the server category. Our performance in the scoring challenge was slightly better, with our simple scoring protocol being the only one capable of identifying an acceptable model for Target 234. This result highlights the robustness of the simple, fully physics-based HADDOCK scoring function, especially when applied to highly flexible antibody-antigen complexes. Inspired by the significant advances in machine learning for structural biology and the dramatic improvement in our success rates after the public release of Alphafold2, we identify the integration of classical approaches like HADDOCK with AI-driven structure prediction methods as a key strategy for improving the accuracy of model generation and scoring.
16 Sep 2024Submitted to PROTEINS: Structure, Function, and Bioinformatics
18 Sep 2024Submission Checks Completed
18 Sep 2024Assigned to Editor
18 Sep 2024Review(s) Completed, Editorial Evaluation Pending
09 Oct 2024Reviewer(s) Assigned
25 Oct 2024Editorial Decision: Revise Minor