MARCEL

Key Contribution

MARCEL provides the first comprehensive benchmark for conformer ensemble learning, demonstrating that explicitly modeling full conformer distributions significantly improves property prediction across drug-like molecules and organometallic catalysts.

Overview

The Molecular Representation and Conformer Ensemble Learning (MARCEL) dataset provides 722K+ conformations across 76K+ molecules spanning four diverse chemical domains: drug-like molecules (Drugs-75K), organophosphorus ligands (Kraken), chiral catalysts (EE), and organometallic complexes (BDE). Unlike prior datasets that focus solely on drug-like molecules, MARCEL enables evaluation of conformer ensemble methods across both pharmaceutical and catalysis applications.

Strengths

• Domain diversity: Beyond drug-like molecules, includes organometallics and catalysts rarely covered in existing benchmarks
• Ensemble-based: Provides full conformer ensembles with statistical weights, not just single conformers
• DFT-quality energies: Drugs-75K features DFT-level conformers and energies (higher accuracy than GEOM-Drugs)
• Realistic scenarios: BDE subset models the practical constraint of lacking DFT-computed conformers for large catalyst systems
• Comprehensive baselines: Benchmarks 18 models across 1D (SMILES), 2D (graph), 3D (single conformer), and ensemble methods
• Property diversity: Covers ionization potential, electron affinity, electronegativity, ligand descriptors, and catalytic properties

Limitations

• Regression only: All tasks are regression; no classification benchmarks
• Chemical space coverage: 76K molecules cannot represent full drug-like or catalyst chemical space
• Compute requirements: Working with large conformer ensembles demands significant computational resources
• Proprietary data: EE subset not publicly available (as of December 2025)
• DFT bottleneck: BDE demonstrates the practical limitation - single DFT optimization can take 2-3 days, making conformer-level DFT infeasible for large organometallics

Technical Notes

Data Generation Pipeline

Drugs-75K

Source: GEOM-Drugs subset

Filtering:

• Minimum 5 rotatable bonds (focus on flexible molecules)
• Allowed elements: H, C, N, O, F, Si, P, S, Cl

Conformer generation:

• DFT-level calculations for both conformers and energies
• Higher accuracy than original GEOM-Drugs (semi-empirical GFN2-xTB)

Properties: Ionization Potential (IP), Electron Affinity (EA), Electronegativity (χ)

Kraken

Source: Original Kraken dataset (1,552 monodentate organophosphorus(III) ligands)

Properties: 5 of 78 available properties

• $B_5$: Ligand sterics (buried volume)
• $L$: Ligand electronics
• $\text{Bur}B_5$: Buried volume variant
• $\text{Bur}L$: Electronic parameter variant

EE (Enantiomeric Excess)

Generation method: Q2MM (Quantum-guided Molecular Mechanics)

Molecules: 872 Rhodium (Rh)-bound atropisomeric catalysts from chiral bisphosphine

Property: Enantiomeric excess (EE) for asymmetric catalysis

Availability: Proprietary-only (not publicly available as of December 2025)

BDE (Bond Dissociation Energy)

Molecules: 5,195 organometallic catalysts (ML₁L₂ structure)

Initial conformers: OpenBabel with geometric optimization

Energies: DFT calculations

Property: Electronic dissociation energy (difference between bound and unbound states)

Key constraint: DFT optimization for full conformer ensembles computationally infeasible (2-3 days per molecule)

Benchmark Setup

Task: Predict molecular properties from structure using different representation strategies (1D/2D/3D/Ensemble)

Data splits: Not explicitly specified in the available information; standard train/validation/test splits used

Hyperparameters: Tuned per model (specific optimization method not detailed in available documentation)

Model categories:

1. 1D Models: SMILES-based (Random Forest on Morgan fingerprints, LSTM, Transformer)
2. 2D Models: Graph-based (GIN, GIN+VN, ChemProp, GraphGPS)
3. 3D Models: Single conformer (SchNet, DimeNet++, GemNet, PaiNN, ClofNet, LEFTNet)
4. Ensemble Models: Full conformer ensemble (same 3D architectures, aggregating over all conformers)

Evaluation metric: Mean Absolute Error (MAE) for all tasks

Key Findings

Ensemble superiority: Across all three benchmarks, ensemble methods (processing full conformer distributions) consistently outperform single-conformer 3D models, with the largest improvements on:

• Drugs-75K: Ensemble GemNet achieves 🥇 0.4066 eV (IP) vs 🥈 0.4069 eV (single conformer)
• Kraken: Ensemble PaiNN achieves 🥇 0.2225 (B₅) vs 0.3443 (single conformer)
• EE: Ensemble GemNet achieves 🥇 11.61% vs 18.03% (single conformer)

3D vs 2D: 3D models generally outperform 2D graph models, especially for conformationally-sensitive properties

Model architecture: GemNet and PaiNN architectures consistently top-ranked across tasks

Dataset Information

Dataset Examples

Dataset Subsets

Related Datasets

Citation

If you use this dataset, please cite:

https://doi.org/10.48550/arXiv.2310.00115