Manufacturing Bits: Nov. 20

Predicting crystal structures
A group of researchers have improved a crystal structure prediction algorithm, enabling the ability to develop new crystal structures and compounds at faster rates.

In 2005, Artem Oganov, now a professor at the Skolkovo Institute of Science and Technology (Skoltech) and the Moscow Institute of Physics and Technology (MIPT), developed a crystal structure prediction algorithm.

This algorithm, called the Universal Structure Predictor: Evolutionary Xtallography (USPEX), can predict a set of structures within a crystal. It can also predict the structures of nanoparticles, polymers and interfaces.

In operation, USPEX determines the composition of the atoms in a crystal. “Then, it generates a small number of random structures whose stability is assessed based on the energy of interaction between the atoms,” according to MIPT. “Next, an evolutionary mechanism starts, where chemists built in natural selection, crossover and mutations of the structures and their ‘descendants’ until they find particularly stable compounds.”

Recently, Skoltech, MIPT and Samara State Technical University improved upon USPEX’s first step and developed a new random structure generator. The generator is based on a database of topological types in crystal structures. In the database, researchers determined that 200,000 inorganic crystal structures belong to 3,000 topological types.

Output of USPEX can be visualised by many codes, e.g. VESTA. USPEX also produces a number of figures in the pdf-format, allowing quick insight into the results. (Source: Skolkovo Institute of Science and Technology)

The new generator can make predications three times faster compared to the previous version. “The 3,000 topological types are the result of abstraction applied to real structures. Going the other way round, you can generate nearly all the known structures and an infinite number of unknown but reasonable structures from these 3,000 types. This is an excellent starting point for an evolutionary mechanism. Right from the start, you most likely sample an area close to the optimal solution. You either get the optimal solution right in the beginning, or get somewhere near it and then get it by evolutionary improvement,” said Pavel Bushlanov, a researcher in the study, on MIPT’s site.

Machine learning for materials
Machine learning has been around for years. This technology makes use of a neural network in a system. In neural networks, the system crunches data and identifies patterns. It matches certain patterns and learns which of those attributes are important.

Machine learning is moving into a number of fields. For example, Virginia Polytechnic Institute and State University (Virginia Tech) has developed a new machine learning framework, enabling new breakthroughs in the design of new materials.

Researchers from Virginia Tech have developed a framework to perform molecular dynamics simulations. This machine learning framework integrates simulations with a particle swarm optimization algorithm and an artificial neural network.

“This novel framework not only uses the machine learning in a unique fashion for the first time, but it also dramatically accelerates the development of accurate computational models of materials,” said Sanket Deshmukh, an assistant professor in the chemical engineering department at Virginia Tech, on the university’s Web site.

“We train the machine learning model in a ‘reverse’ fashion by using the properties of a model obtained from molecular dynamics simulations as an input for the machine learning model, and using the input parameters used in molecular dynamics simulations as an output for the machine learning model,” said Karteek Bejagam, a post-doctoral researcher at Virginia Tech.

Finding enzymes
The University of Oxford has found a way of predicting enzyme activity in living organisms.

Enzymes are protein catalysts. They convert and accelerate substances into different molecules. If researchers can predict the functions of enzymes, they can impact the behavior of them and develop drugs to treat disease.

Researchers from Oxford looked at enzymes from a plant species. Using techniques from Oxford’s machine learning group, researchers developed a way to predict enzymes.

“The key thing is that rather than being a ‘black box,’ this method gives back to the chemist/biologist successful predictions and reasons for those predictions that have chemical and biological meaning. This in turn has allowed us to work out which enzymes can be used in synthesis, predict the activity of enzymes from very different species (even bacteria) and to work out how to engineer enzymes in a new way based on suggestions that we wouldn’t have predicted,” said Ben Davis, professor of chemistry at the University of Oxford, on the university’s Web site.