GDB-BSIE: Small Molecule Quantum Energy Calculations

A dataset of 13,221 small molecules with up to 7 heavy atoms derived from the GDB-11 and GDB-13 chemical universe databases. For each molecule, it includes a minimized geometry and up to 10 geometries sampled from molecular dynamics, with RHF and B3LYP energies computed using five basis sets in the Psi4 package. The dataset was created by Søren Holm of Stanford University.

Use Cases

• Training machine learning models for quantum energy prediction based on molecular geometry.
• Benchmarking computational chemistry methods using the provided RHF and B3LYP energy calculations.
• Studying the relationship between molecular structure and electronic energy across different basis sets.
• Developing models for molecular dynamics simulation using the sampled geometry data.

Strengths

• Contains data for 13,221 unique small molecules.
• Provides multiple energy calculations per molecule using two methods (RHF, B3LYP) and five basis sets.
• Includes multiple molecular geometries per molecule: one minimized and up to ten sampled from dynamics.

Limitations

• Row count for the full dataset of geometries and energies is unknown, which may limit suitability assessment.
• Column-level documentation is absent; field semantics must be inferred after download.
• Last update date is unknown; freshness unverified.

Provenance

Source

Derived from the GDB-11 and GDB-13 chemical universe databases.

Collection Method

Molecular geometries were minimized and sampled from MD; energies were computed using the Psi4 computational chemistry package.