← Research index

Research story 01 / CrystEngComm · 2024

Which crystals are worth a closer look?

Learning where to spend the next quantum calculation.

Read the paper ↗Explore the dataset ↗
Fig. 01 / Crystal energiesFentanyl
Force-field energy versus relative DFT energy1780 deposited records. Each point pairs the initial force-field energy with the energy after DFT relaxation. Orange points are within 1 kcal per mole of the lowest deposited DFT energy.03601.252.5Relative DFT energy / kcal mol⁻¹Force-field energy / kcal mol⁻¹The low-energyregion of interest
Within 1 kcal/mol1,780 records shown
Each point pairs a generated packing with a DFT relaxation. DFT energies are relative to the lowest deposited value. Published data ↗
01

The question

A molecule can pack into many different crystals. Finding promising solid forms requires both a broad search and accurate energy estimates. The expensive step is relaxing candidate structures with density functional theory (DFT). Which candidates should receive that calculation next?

02

The approach

Our model learns the energy correction between force-field minima and DFT-relaxed structures. SOAP representations and a committee of ridge-regression models provide predictions with uncertainty. An active-learning loop selects candidates, calculates their DFT energies, and updates the model until a chosen convergence threshold is reached.

03

What we found

In the fentanyl benchmark, the paper reports 34% savings in reranking computational cost and recovery of 173 of the 174 structures in the target 1 kcal/mol window. A further study examined 17 pharmaceutical compounds using preselected structure pools.

04

Where the result applies

The method reranks an existing pool; it cannot recover a packing that was never generated. The study considers zero-kelvin energies, without finite-temperature sampling. The 17-compound exercise uses reduced pools and is not a direct comparison with a complete GRACE search.

Paper & resources

A. Anelli, H. Dietrich, P. Ectors, F. Stowasser, T. Bereau, M. Neumann and J. van den Ende. Robust and efficient reranking in crystal structure prediction: a data driven method for real-life molecules. CrystEngComm, 2024, 26, 5845–5849.

DOI: 10.1039/D4CE00752B ↗Fentanyl data on Zenodo ↗

The publication’s implementation and internal compound landscapes are proprietary. The fentanyl data are publicly available.