Learning Disentangled Equivariant Representation for Explicitly Controllable 3D Molecule Generation

Authors: Haoran Liu, Youzhi Luo, Tianxiao Li, James Caverlee, Martin Renqiang Min

AAAI 2025 | Venue PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental Extensive experiments validate our model s effectiveness on property-guided and context-guided molecule generation, both for de-novo 3D molecule design and structure-based drug discovery against protein targets.
Researcher Affiliation Collaboration 1Texas A&M University, College Station, TX 77840; 2NEC Laboratories America, Princeton, NJ 08540
Pseudocode Yes A detailed algorithm for this decoding process is in Appendix 9.
Open Source Code No Supplementary materials are available at this link1. 1https://arxiv.org/abs/2412.15086
Open Datasets Yes We conduct experiments mainly on two benchmark datasets for 3D molecule generation: GEOM-Drugs (Axelrod and Gomez-Bombarelli 2022) and Cross Docked2020 (Francoeur et al. 2020).
Dataset Splits No Our model is trained using a subset of 50,000 molecular structures from each dataset, specifically selecting those with the lowest energy conformations for each molecule.
Hardware Specification No The paper does not provide specific hardware details such as exact GPU/CPU models, processor types with speeds, or memory amounts used for running its experiments.
Software Dependencies No The paper mentions 'RDKit (Landrum 2016)' but does not provide a comprehensive list of key software components with their specific version numbers needed to replicate the experiments.
Experiment Setup No The paper states 'Implementation details are provided in Appendix 11.' but the main text does not contain specific hyperparameter values, training configurations, or system-level settings for the experiments.