MolMiner: Toward Controllable, Three‐Dimensional‐Aware, Fragment‐Based Molecular Design
- 1Department of Energy Conversion and Storage CAPeX Pioneer Center for Accelerating P2X Materials Discovery, Technical University of Denmark Lyngby Denmark
- 2Department of Applied Mathematics and Computer Science Technical University of Denmark Kongens Lyngby Denmark
Abstract
We introduce MolMiner, a fragment‐based, geometry‐aware, and order‐agnostic autoregressive model for molecular design. MolMiner supports high‐dimensional conditional control over 12 physicochemical and structural properties from partial specifications, constructs molecules via symmetry‐aware fragment attachments, and conditions each generation step on force‐field‐relaxed three‐dimensional geometry of the partial structure. Conditional control emerges without auxiliary property losses. On targeted property windows, conditioning lifts hit rates by up to 5.25 over unconditional generation and 3.5 over the training distribution itself—overriding the model's intrinsic biases—at the cost of a small reduction in unconditional distributional fidelity. MolMiner unifies dynamic geometry, symmetry handling, order‐agnostic generation, and scalable multi‐property conditioning within a single framework.
Related
- GCMDiff: a condition-guided diffusion framework for pharmacophore-fragment drug discovery and target-specific molecular design
- PyFMO: A Python Framework for Fragment Molecular Orbital Calculations
- DeepSAP: an integrated generative and predictive deep-learning framework for controllable design of self-assembling peptides.
- Conditional latent distillation enables controllable molecular generation in data-limited soft materials
- Transformer-based molecular fragment prediction using SMILES and DeepSMILES representations in a fragment-based drug discovery pipeline.
BibTeX
@article{ochoa2026molminer,
title = {{MolMiner: Toward Controllable, Three‐Dimensional‐Aware, Fragment‐Based Molecular Design}},
author = {Raul Ortega Ochoa and Tejs Vegge and Jes Frellsen},
journal = {Advanced Intelligent Discovery},
year = {2026},
month = jul,
doi = {10.1002/aidi.70138},
url = {https://doi.org/10.1002/aidi.70138},
note = {CC BY},
}
RIS
TY - JOUR
TI - MolMiner: Toward Controllable, Three‐Dimensional‐Aware, Fragment‐Based Molecular Design
AU - Raul Ortega Ochoa
AU - Tejs Vegge
AU - Jes Frellsen
PY - 2026
DA - 2026/07/06
JO - Advanced Intelligent Discovery
DO - 10.1002/aidi.70138
UR - https://doi.org/10.1002/aidi.70138
AB - We introduce MolMiner, a fragment‐based, geometry‐aware, and order‐agnostic autoregressive model for molecular design. MolMiner supports high‐dimensional conditional control over 12 physicochemical and structural properties from partial specifications, constructs molecules via symmetry‐aware fragment attachments, and conditions each generation step on force‐field‐relaxed three‐dimensional geometry of the partial structure. Conditional control emerges without auxiliary property losses. On targeted property windows, conditioning lifts hit rates by up to 5.25 over unconditional generation and 3.5 over the training distribution itself—overriding the model's intrinsic biases—at the cost of a small reduction in unconditional distributional fidelity. MolMiner unifies dynamic geometry, symmetry handling, order‐agnostic generation, and scalable multi‐property conditioning within a single framework.
SN - 2943-9981
ER -