Skip to content
DraconiticResearch
Search Sources API
197,768 studies31 sourcesUpdated 57 min ago
Peer-reviewedCC BYBioinformatics (Oxford, England)

Pocket contrastive learning and conformation fusion for robust structure-conditioned 3D molecular generation under receptor conformational variation.

  • Huan Liu1
  • Yijie Ding2
  • Xiaoyi Guo3
  • Hongjie Wu1
  1. 1College of Electronic and Information Engineering, Suzhou University of Science and Technology, Suzhou, 215009, China.
  2. 2Faculty of Computing, Harbin Institute of Technology, Harbin, 150001, China.
  3. 3Li Ka Shing Faculty of Medicine, The University of Hong Kong, Hong Kong SAR, China.
Read full textPDFDOICite

Abstract

Motivation

Pocket-conditioned molecular generation incorporates the three-dimensional (3D) geometry and chemical environment of protein binding sites into molecular design. However, receptor conformational variation can introduce distribution shifts between training and inference, particularly under cross-docking conditions. Existing approaches often do not fully exploit the correspondence between related receptor conformations during training.Results

We propose a structure-conditioned 3D molecular generation framework that combines pocket-level contrastive learning with reference-conformation feature fusion. Built on a stepwise constructive generative model, these training-stage components encourage stable pocket representations across related conformations while preserving single-pocket inference without requiring a reference conformation. Evaluation on CrossDocked2020 and three external datasets, ASB-E, DUD-E, and APObind, demonstrates improvements in generation success, structural validity, and docking-related performance, particularly under cross-dock evaluation. These results support improved robustness of pocket-conditioned molecular generation under receptor conformational variation.Availability and implementation

The source code and reproduction instructions are available at https://github.com/LHHHHuan/MTMGen. Pretrained models and an archival code snapshot are available at https://doi.org/10.5281/zenodo.21944857.

Related

  1. TopoFlow: Evolutionarily Conditioned Flow Matching for Protein Conformational Ensemble Generation
    X. Yang, X. Cui, X. LI +4
    bioRxivPreprintAll rights reservedFull textBioinformatics & computational biologybioRxiv
  2. MIEN: a deep-learning model for pocket-conditioned glycan pose and conformation discrimination
    X. Huang, Z. Tian
    bioRxivPreprintAll rights reservedFull textBioinformatics & computational biologybioRxiv
  3. Growth, conformation, and egg-weight traits of CARI Nirbheek and CARI Shyama chickens across the foundation cross and first inter se generation.
    Jaydip Rokade, Roshni Chand, Nitish Gaitri +5
    Frontiers in veterinary sciencePeer-reviewedCC BYFull textZoology & animal behaviourEurope PMC
  4. Pocket-Conditioned Generation and Down-Selection of Non-Oxime Reactivators
    D. Jimenez, W. K. Clay, C. Barajas
    bioRxivPreprintAll rights reservedFull textSystems & synthetic biologybioRxiv
  5. trFlow: ultrafast all-atom protein conformational ensemble generation via conditional flow matching
    C. Xiang, K. Zhao, Z. Peng +2
    bioRxivPreprintCC BY-NC-NDFull textBioinformatics & computational biologybioRxiv

Cite

BibTeXRIS
BibTeX
@article{liu2026pocket,
  title = {{Pocket contrastive learning and conformation fusion for robust structure-conditioned 3D molecular generation under receptor conformational variation.}},
  author = {Huan Liu and Yijie Ding and Xiaoyi Guo and Hongjie Wu},
  journal = {Bioinformatics (Oxford, England)},
  year = {2026},
  month = oct,
  doi = {10.1093/bioinformatics/btag740},
  url = {https://europepmc.org/article/MED/42834449},
  note = {CC BY},
}
RIS
TY  - JOUR
TI  - Pocket contrastive learning and conformation fusion for robust structure-conditioned 3D molecular generation under receptor conformational variation.
AU  - Huan Liu
AU  - Yijie Ding
AU  - Xiaoyi Guo
AU  - Hongjie Wu
PY  - 2026
DA  - 2026/10/01
JO  - Bioinformatics (Oxford, England)
DO  - 10.1093/bioinformatics/btag740
UR  - https://europepmc.org/article/MED/42834449
AB  - Motivation  Pocket-conditioned molecular generation incorporates the three-dimensional (3D) geometry and chemical environment of protein binding sites into molecular design. However, receptor conformational variation can introduce distribution shifts between training and inference, particularly under cross-docking conditions. Existing approaches often do not fully exploit the correspondence between related receptor conformations during training.Results  We propose a structure-conditioned 3D molecular generation framework that combines pocket-level contrastive learning with reference-conformation feature fusion. Built on a stepwise constructive generative model, these training-stage components encourage stable pocket representations across related conformations while preserving single-pocket inference without requiring a reference conformation. Evaluation on CrossDocked2020 and three external datasets, ASB-E, DUD-E, and APObind, demonstrates improvements in generation success, structural validity, and docking-related performance, particularly under cross-dock evaluation. These results support improved robustness of pocket-conditioned molecular generation under receptor conformational variation.Availability and implementation  The source code and reproduction instructions are available at https://github.com/LHHHHuan/MTMGen. Pretrained models and an archival code snapshot are available at https://doi.org/10.5281/zenodo.21944857.
SN  - 1367-4803
SN  - 1367-4811
ER  -