CrystalConf 2023

Computational chemistry is a rapidly growing field that uses computers for running simulations to study the properties of molecules and materials. Computer-aided drug design (CADD) is a branch of computational chemistry that uses molecular modelling and simulation to help understand and design new drugs.

Scientists often need to deal with tons of data (big data) generated by specialized software, which needs to be processed to extract relevant information. We have developed a modern computational chemistry library, termed chem.cr, for data manipulation and analysis using the Crystal language.

chem.cr aims to be both fast and easy to use, and it currently provides several features:

• Hierarchical object-oriented access to molecular structure
• Iterator-based file reading of several formats (PDB, Mol2, etc)
• Topology detection
• Spatial measurements
• Volumetric data
• Type safety
• Fast performance, rivaling or even surpassing other libraries built with Python/C/C++

chem.cr has already been used to develop and publish a new algorithm for protein structure (Adasme-Carreño et al., 2021), and we're developing new protocols for molecular docking and free energy calculations (Femdock and Moltiverse) that are crucial for CADD.

• Femdock helps to sample the possible orientations of a drug within a protein binding site using a genetic algorithm, like other molecular docking approaches, but thanks to the Crystal language, the code is clearer and easier to understand and modify.
• Moltiverse is used to generate molecular conformers of the drug using enhanced sampling methods, that directly support the exploration performed by Femdock.

We hope that this new library and protocols will help scientists to develop new/better tools that accelerate drug discovery and the study of molecular structure.

References

• Adasme-Carreño, F., Caballero, J., & Ireta, J. (2021). PSIQUE: Protein Secondary Structure Identification on the Basis of Quaternions and Electronic Structure Calculations. Journal of Chemical Information and Modeling, 61(4), 1789–1800. https://doi.org/10.1021/acs.jcim.0c01343