A Cross-Domain Graph Learning Protocol for Single-Step Molecular Geometry Refinement

2026-01-30

AI Summary

This work addresses the high computational cost of DFT-based geometry optimization in high-throughput molecular screening by proposing GeoOpt-Net, a multi-branch SE(3)-equivariant graph neural network that directly refines force-field initial conformations to B3LYP/TZVP-level accuracy in a single forward pass. The method introduces a fidelity-aware feature modulation (FAFM) mechanism and employs a two-stage training strategy to achieve high geometric and energetic consistency across diverse theoretical levels and basis sets. Evaluated on drug-like molecules, GeoOpt-Net attains sub-milliangstrom all-atom RMSD and near-zero single-point energy deviations, with a DFT convergence rate of 65.0% under relaxed criteria, substantially reducing the need for subsequent re-optimization and associated computational overhead.

Technology Category

• Search and Optimization: Learning to Search
• Machine Learning: Optimization
• Constraint Satisfaction and Optimization: Other Foundations of Constraint Satisfaction

Application Category

• Graph Algorithms and Modeling for the Web: Graph neural networks and deep learning approaches for Web-related graphs
• Economics, Online Markets and Human Computation: Trust and reliance of crowd workers and data experts on GenAI
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking

Abstract

Accurate molecular geometries are a prerequisite for reliable quantum-chemical predictions, yet density functional theory (DFT) optimization remains a major bottleneck for high-throughput molecular screening. Here we present GeoOpt-Net, a multi-branch SE(3)-equivariant geometry refinement network that predicts DFT-quality structures at the B3LYP/TZVP level of theory in a single forward pass starting from inexpensive initial conformers generated at a low-cost force-field level. GeoOpt-Net is trained using a two-stage strategy in which a broadly pretrained geometric representation is subsequently fine-tuned to approach B3LYP/TZVP-level accuracy, with theory- and basis-set-aware calibration enabled by a fidelity-aware feature modulation (FAFM) mechanism. Benchmarking against representative approaches spanning classical conformer generation (RDKit), semiempirical quantum methods (xTB), data-driven geometry refinement pipelines (Auto3D), and machine-learning interatomic potentials (UMA) on external drug-like molecules demonstrates that GeoOpt-Net achieves sub-milli-\AA{} all-atom RMSD with near-zero B3LYP/TZVP single-point energy deviations, indicating DFT-ready geometries that closely reproduce both structural and energetic references. Beyond geometric metrics, GeoOpt-Net generates initial guesses intrinsically compatible with DFT convergence criteria, yielding nonzero ``All-YES''convergence rates (65.0\% under loose and 33.4\% under default thresholds), and substantially reducing re-optimization steps and wall-clock time. GeoOpt-Net further exhibits smooth and predictable energy scaling with molecular complexity while preserving key electronic observables such as dipole moments. Collectively, these results establish GeoOpt-Net as a scalable, physically consistent geometry refinement framework that enables efficient acceleration of DFT-based quantum-chemical workflows.

Problem

• molecular geometry refinement
• density functional theory
• high-throughput screening
• computational bottleneck
• DFT-quality structures

Innovation

• SE(3)-equivariant
• geometry refinement
• fidelity-aware feature modulation
• DFT acceleration
• molecular conformation