A scalable and quantum-accurate foundation model for biomolecular force fields via linearly tensorized quadrangle attention

Physics-informed deep learning: A new paradigm for macromolecular recognition and rational protein design

By delineating how physics-based priors synergize with data-driven representation learning, this review provides a comprehensive roadmap for generating physically plausible and thermodynamically stable therapeutics, ultimately accelerating the transition of computationally designed molecules from in silico blueprints to viable clinical candidates.

Modeling dual-range atomic interactions with physicochemical principles for molecular force fields

GeoNet is a physicochemical-principle-guided framework for modeling dual-range atomic interactions that achieves the smallest model size and the shortest training time, demonstrating both superior predictive performance and computational efficiency.

Efficient Force Evaluation via Fragmentation: Toward Geometry Optimization of Protein-Ligand Systems with Quantum Mechanical Accuracy.

This work presents a hybrid fragmentation-based approach, FragQMMM, that enables efficient and accurate force evaluation for large molecular systems, and successfully reproduces the hydrogen-bond network and preserving the underlying structure-activity relationship.

Molecular physics-informed neural network (mPINN) for solving the molecular dynamics equation of motion with energy conservation

The results demonstrate that the mPINN architecture functions as a reliable, physics-constrained ML framework capable of delivering high-fidelity trajectory predictions for complex multi-body molecular systems.

HessFit: Deriving Molecule-Specific Force Fields from Quantum Mechanical Hessians.

OmniScore: Universal Scoring of Diverse Biomolecular Complexes via Equivariant Geometry-Aware Discrete Representation Learning

Scoring biomolecular complexes is central to structure assessment and drug discovery, yet the complexes themselves vary widely in pose, size, and molecular composition. A scoring function tuned for one interaction type rarely carries over to another, and most existing methods compound the problem by leaning heavily on task-specific labels. We introduce OmniScore, a universal structure-based framework that learns a shared geometry-aware representation of complexes once and then adapts it to downstream scoring through lightweight task-specific heads. OmniScore couples a graph view and a sequence view of each structure, encodes its three-dimensional geometry, and compresses representations into a compact latent space that a reconstruction module and prediction heads can reuse. We pretrain this backbone on diverse datasets including complexes, monomers, and small molecules with complementary objectives: coordinate recovery, correcting corrupted input tokens, predicting molecular identity, and grounding the representation in structure-level physical quantities. Across the evaluated benchmarks, OmniScore gave the best antibody-antigen and nanobody-antigen quality assessment on all reported metrics compared to state-of-the-art baselines. Its frozen residue embeddings matched the state-of-the-art protein-tokenization method with an average functional-site accuracy of 71.8% on a standard residue-level benchmark. On protein-ligand scoring and ranking benchmarks, it performed on par with methods built specifically for that single task. These results suggest that geometry-aware pretraining can provide a reusable scoring backbone for tasks that depend on interfacial and residue-level structure, within the evaluated settings.