CREST—A program for the exploration of low-energy molecular chemical space

CREST—A program for the exploration of low-energy molecular chemical space

March 21, 2024 | Philipp Pracht; Stefan Grimme; Christoph Bannwarth; Fabian Bohle; Sebastian Ehler; Gereon Feldmann; Johannes Gorges; Marcel Müller; Tim Neudecker; Christoph Plett; Sebastian Spicher; Pit Steinbach; Patryk A. Wesolowski; Felix Zeller
CREST is an open-source program for efficient and automated exploration of molecular chemical space. Originally developed as an automated driver for extended tight-binding (xTB) calculations, it offers capabilities for conformational sampling, geometry optimization, and molecular structure analysis. The program includes automated procedures for conformational sampling, explicit solvation studies, and identification of protonation/deprotonation sites. Calculations are set up for concurrent execution, enabling efficient single-node parallelization. CREST requires minimal user input and includes implementations of GFNn-xTB Hamiltonians and GFN-FF force-field. Interfaces to quantum chemistry and force-field software can be easily created. Recent developments in CREST 3.0 include a refactored calculation backend, providing significant speed-up for sampling small or medium-sized drug molecules and allowing more sophisticated setups, such as quantum mechanics/molecular mechanics and minimum energy crossing point calculations. CREST 3.0 is more suitable for high-throughput conformational sampling of small-molecule databases than previous versions. The program uses semiempirical methods for conformational sampling, which are computationally efficient and accurate for exploring chemical space. CREST 3.0 has been benchmarked against various methods, showing good performance for conformational energy calculations. The program also includes an ensemble sorting algorithm to identify conformers, rotamers, and duplicates. CREST 3.0 can model explicit and micro-solvation using the Quantum Cluster Growth (QCG) algorithm, allowing automated generation of cluster ensembles containing solvated solutes. The QCG algorithm includes two essential steps: growth of a low-energy solute-solvent cluster and generation of different conformers from this cluster. The algorithm uses wall potentials to ensure even distribution of solvent molecules around the solute. CREST 3.0 can compute solvation free energies from generated ensembles. The program is designed to be modular and interoperable, enabling seamless integration with other software for chemical physics research.CREST is an open-source program for efficient and automated exploration of molecular chemical space. Originally developed as an automated driver for extended tight-binding (xTB) calculations, it offers capabilities for conformational sampling, geometry optimization, and molecular structure analysis. The program includes automated procedures for conformational sampling, explicit solvation studies, and identification of protonation/deprotonation sites. Calculations are set up for concurrent execution, enabling efficient single-node parallelization. CREST requires minimal user input and includes implementations of GFNn-xTB Hamiltonians and GFN-FF force-field. Interfaces to quantum chemistry and force-field software can be easily created. Recent developments in CREST 3.0 include a refactored calculation backend, providing significant speed-up for sampling small or medium-sized drug molecules and allowing more sophisticated setups, such as quantum mechanics/molecular mechanics and minimum energy crossing point calculations. CREST 3.0 is more suitable for high-throughput conformational sampling of small-molecule databases than previous versions. The program uses semiempirical methods for conformational sampling, which are computationally efficient and accurate for exploring chemical space. CREST 3.0 has been benchmarked against various methods, showing good performance for conformational energy calculations. The program also includes an ensemble sorting algorithm to identify conformers, rotamers, and duplicates. CREST 3.0 can model explicit and micro-solvation using the Quantum Cluster Growth (QCG) algorithm, allowing automated generation of cluster ensembles containing solvated solutes. The QCG algorithm includes two essential steps: growth of a low-energy solute-solvent cluster and generation of different conformers from this cluster. The algorithm uses wall potentials to ensure even distribution of solvent molecules around the solute. CREST 3.0 can compute solvation free energies from generated ensembles. The program is designed to be modular and interoperable, enabling seamless integration with other software for chemical physics research.
Reach us at info@study.space
Understanding CREST-A program for the exploration of low-energy molecular chemical space.