Learning Molecular Chirality via Chiral Determinant Kernels
Organizations: AGI Institute, School of Computer Science, Shanghai Jiao Tong University · Shanghai Innovation Institute
Abstract
Chirality is a fundamental molecular property that governs stereospecific behavior in chemistry and biology. Capturing chirality in machine learning models remains challenging due to the geometric complexity of stereochemical relationships and the limitations of traditional molecular representations that often lack explicit stereochemical encoding. Existing approaches to chiral molecular representation primarily focus on central chirality, relying on handcrafted stereochemical tags or limited 3D encodings, and thus fail to generalize to more complex forms such as axial chirality. In this work, we introduce ChiDeK (Chiral Determinant Kernels), a framework that systematically integrates stereogenic information into molecular representation learning. We propose the chiral determinant kernel to encode the SE(3)-invariant chirality product and employ cross-attention to integrate stereochemical information from local chiral centers into the global molecular representation. This design enables explicit modeling of chiral-related features within a unified architecture, capable of jointly encoding central and axial chirality. To support the evaluation of axial chirality, we construct a new benchmark for electronic circular dichroism (ECD) and optical rotation (OR) prediction. Across four tasks, including R/S configuration classification, enantiomer ranking, ECD spectrum prediction, and OR prediction, ChiDeK achieves substantial improvements over state-of-the-art baselines, most notably yielding over 7% higher accuracy on axially chiral tasks on average.
Figures & tables
| Method | R/S (%) | Ranking (%) | Position | Number | Symbol (%) |
|---|---|---|---|---|---|
| Acc | Acc | RMSE | RMSE | Acc | |
| DimeNet++ | |||||
| Tetra-DMPNN (c) | |||||
| Tetra-DMPNN (p) | |||||
| SphereNet | |||||
| ChIRo |
| Method | Rotation (%) | Position | Number | Symbol (%) |
|---|---|---|---|---|
| Acc | RMSE | RMSE | Acc | |
| DimeNet++ | ||||
| Tetra-DMPNN (c) | 3.15 0.12 | |||
| Tetra-DMPNN (p) | 1.05 0.12 | |||
| SphereNet | ||||
| ChIRo |
| Model Components | Acc (%) | ||
|---|---|---|---|
| Chiral Encoder | Rank strategy | Separate | |
| Linear w/o. | None | ✓ | |
| Linear w. | None | ✓ | |
| Kernel-based | QR | ✗ | |
| Kernel-based | Reg | ✓ | |
| Kernel-based | None | ✓ | |
Appendix figures & tables13 assets
Supplementary material from the paper’s appendix.
Appendix
| Dataset | Chirality | Task | Size | Split |
|---|---|---|---|---|
| ChIRo (R/S) | Central | R/S classification | 466K conformers / 78K mols | 70/15/15 |
| ChIRo (Ranking) | Central | Enantiomer ranking | 335K conformers / 69K mols | 70/15/15 |
| CMCDS | Central | ECD prediction | 22,182 conformers / mols | 80/10/10 |
| ACMP | Axial | ECD and OR prediction | 1,192 conformers / mols | 80/10/10 |
| Subtype | # Enantiomers | # Height symbols | Acc (%) |
|---|---|---|---|
| Chiral atom pair | 10 | 26 | |
| Allene-like | 2 | 4 | |
| Spiral atom and chain | 4 | 8 | |
| Biaryl | 70 | 194 | |
| Heterobiaryl (C-N) | 12 | 20 | |
| Heterobiaryl (C-B) | 6 | 16 |
| Metric | ROC-AUC (%) | RMSE | |||
|---|---|---|---|---|---|
| Datasets | BBBP | BACE | ClinTox | SIDER | FreeSolv |
| # Molecules | 2039 | 1513 | 1478 | 1427 | 642 |
| # Tasks | 1 | 1 | 2 | 27 | 1 |
| W/o. pre-training | |||||
| D-MPNN | |||||
| Attentive FP | |||||
| Ratio | W/o. noise | 2.5% | 5% | 7.5% | 10% |
|---|---|---|---|---|---|
| ChiDeK |
| Ratio | W/o. noise | W. noise |
|---|---|---|
| ChiDeK |
| Method | Rotation (%) | Symbol (%) |
|---|---|---|
| ChIRo | ||
| ChIRo + Chirality matrix | ||
| ChiDeK |
| Hyperparameter | Search Space |
|---|---|
| # Chiral Transformer Layers | |
| # Chiral Transformer Heads | |
| Hidden Dimension | |
| Projection Dimension | |
| Rank Strategy | |
| Learning Rate | {1e-4,5e-4} |
Molecular Representation Learning
Molecular representation learning encodes chemical structures and context for prediction, generation, and discovery across chemistry and biology.
See all 219 papers