Molecular crystal structures pack it up

Whether organic chemists are working on developing new molecular energy or creating new barrier drugs in the pharmaceutical industry, each one is looking at how you can maximize the molecular chemical structure to the target features you want to achieve.

Part of that optimization involves the molecular crystal packaging motif, a pattern that is seen in terms of how molecules move relative to each other within a crystal structure. The conventional packaging motor databases have remained small due to intensive manual labeling processes and inadequate labeling schemes.

To help solve this problem, a team of Lawrence Livermore National Laboratory (LLNL) materials and computer scientists has developed a freely available package, Autopack, which formalizes the packaging motif labeling process and automated processing and labeling could be the packaging badges of thousands of molecular crystals. structur. The research appears in the Journal of Chemical Information and Modeling.

Small-scale crystal engineering studies over the past 30 years suggest that, although they predict experimental crystal structures from a chemical structure on their own out of reach, relationships may exist between molecular chemical structures and a special feature of the crystal structure called the packing motif.

Molecular crystal packaging motif is an important concept for athletics and organic electronics applications due to a known correlation between molecular crystal packaging motifs and interesting performance characteristics, which includes instability for molecular explosives and cost transport for molecular semiconductors.

To date, a formal and open source method of packaging branding has never been specified. Instead, packaging motifs are characterized to direct molecular crystals by human evaluation of crystal structure and diagnosis, leading to small and acoustic data sets.

“In the age of machine learning, the ability to create large databases, with labels of molecular crystal packaging motifs is now particularly important,” said LLNL data scientist Donald Loveland, lead author of the paper. “Efforts like this could generate models that can predict packaging motifs from a molecular chemical structure alone, which would help organic chemists prioritize the synthesis of new molecules based on the packaging motor and the features that are missing. “

The new LLNL operation uses an efficient optimization algorithm that circumvents many problems found in previously proposed packaging motif labeling methods, leading to new advanced results when tested on a currant dataset. LLNL.

Through Autopack, researchers have been able to generate a data set of nearly 10,000 packaging motifs for a set of energetic and energetic molecules of interest to the Lab, a task that would have been impossible before. For context, previous literature has captured the order of 100 molecules due to their boring and timeless nature of manual labeling. Early analysis of this new data set notes complex shifts between intermolecular interactions, 3-D molecular coherence and packaging motifs that have not been studied in the field at present. , providing guidance on the next steps for crystal engineering pipes.

The code is freely available through the Office of Innovations and Partnerships Lab.