Hexbyte Glen Cove New strategy puts evolution of microscopic structures on fast track thumbnail

Hexbyte Glen Cove New strategy puts evolution of microscopic structures on fast track

Hexbyte Glen Cove

Engineers at Rice University and Lawrence Livermore National Laboratory are using neural networks to accelerate the prediction of how microstructures of materials evolve. This example predicts snowflake-like dendritic crystal growth. Credit: Mesoscale Materials Science Group/Rice University

The microscopic structures and properties of materials are intimately linked, and customizing them is a challenge. Rice University engineers are determined to simplify the process through machine learning.

To that end, the Rice lab of materials scientist Ming Tang, in collaboration with physicist Fei Zhou at Lawrence Livermore National Laboratory, introduced a technique to predict the evolution of microstructures—structural features between 10 nanometers and 100 microns—in materials.

Their open-access paper in the Cell Press journal Patterns shows how (computer models that mimic the brain’s neurons) can train themselves to predict how a structure will grow under a certain environment, much like a snowflake forms from moisture in nature.

In fact, snowflake-like, dendritic crystal structures were one of the examples the lab used in its proof-of-concept study.

“In modern material science, it’s widely accepted that the microstructure often plays a critical role in controlling a material’s properties,” Tang said. “You not only want to control how the atoms are arranged on lattices, but also what the microstructure looks like, to give you good performance and even new functionality.

“The holy grail of designing materials is to be able to predict how a microstructure will change under given conditions, whether we heat it up or apply stress or some other type of stimulation,” he said.

Tang has worked to refine microstructure prediction for his entire career, but said the traditional equation-based approach faces significant challenges to allow scientists to keep up with the demand for new materials.

“The tremendous progress in machine learning encouraged Fei at Lawrence Livermore and us to see if we could apply it to materials,” he said.






Fortunately, there was plenty of data from the traditional method to help train the team’s neural networks, which view the early evolution of microstructures to predict the next step, and the next one, and so on.

“This is what machinery is good at, seeing the correlation in a very complex way that the human mind is not able to,” Tang said. “We take advantage of that.”

The researchers tested their neural networks on four distinct types of microstructure: plane-wave propagation, grain growth, spinodal decomposition and dendritic crystal growth.

In each test, the networks were fed between 1,000 and 2,000 sets of 20 successive images illustrating a material’s microstructure evolution as predicted by the equations. After learning the evolution rules from these data, the was then given from 1 to 10 images to predict the next 50 to 200 frames, and usually did so in seconds.

The new technique’s advantages quickly became clear: The neural networks, powered by graphic processors, sped the computations up to 718 times for grain growth, compared to the previous algorithm. When run on a standard central processor, they were still up to 87 times faster than the old method. The prediction of other types of evolution showed similar, though not as dramatic, speed increases.

Comparisons with images from the traditional simulation method proved the predictions were largely on the mark, Tang said. “Based on that, we see how we can update the parameters to make the prediction more and more accurate,” he said. “Then we can use these predictions to help design materials we have not seen before.

“Another benefit is that it’s able to make predictions even when we do not know everything about the material properties in a system,” Tang said. “We couldn’t do that with the equation-based method, which needs to know all the parameter values in the equations to perform simulations.”

Tang said the computation efficiency of neural networks could accelerate the development of novel materials. He expects that will be helpful in his lab’s ongoing design of more efficient batteries. “We’re thinking about novel three-dimensional structures that will help charge and discharge batteries much faster than what we have now,” Tang said. “This is an that is perfect for our new approach.”



More information:
Kaiqi Yang et al, Self-supervised learning and prediction of microstructure evolution with convolutional recurrent neural networks, Patterns (2021). DOI: 10.1016/j.patter.2021.100243

Citation:
New strategy puts evolution of microscopic structures on fast track (2021, April 30)
retrieved 2 May 2021
from https://phys.org/news/2021-04-strategy-evolution-microscopic-fast-track.html

This document is subject to copyright. Apart from any fair dealing for the purpose of private study or research, no
part may be reproduced without the written permission. The content is provided for information purposes only.

Read More Hexbyte Glen Cove Educational Blog Repost With Backlinks —

Hexbyte Glen Cove Two-birds-one-stone strategy shows promise in RNA-repeat expansion diseases thumbnail

Hexbyte Glen Cove Two-birds-one-stone strategy shows promise in RNA-repeat expansion diseases

Hexbyte Glen Cove

Alicia Angelbello and Matthew Disney, PhD, in the Disney lab on Scripps Research’s Jupiter, Florida, campus. Credit: The Scripps Research Institute

A new strategy for treating a variety of diseases known as RNA-repeat expansion disorders, which affect millions of people, has shown promise in proof-of-principle tests conducted by scientists at Scripps Research.

The results suggest that someday, a handful of well-targeted drugs might be able to treat the more than 40 human disorders—including Huntington’s and variants of amyotrophic lateral sclerosis (ALS)—that arise from RNA-repeat expansions.

“This study lays a foundation for the development of drugs that can address multiple repeat-expansion diseases by targeting shared abnormal structures on their RNAs,” says the study’s principal investigator Matthew Disney, Ph.D., professor of chemistry at Scripps Research.

In RNA-repeat expansion diseases, contain excess DNA in the form of dozens or even hundreds of repeating short strings of DNA “letters.” In where these mutant genes are active, that DNA is copied out into RNA on the way to being translated into proteins. The resulting abnormal RNAs can cause trouble in a variety of ways, such as by folding up into structures that are toxic to cells.

In the study, published in Cell Chemical Biology, the scientists showed that a potential drug molecule they developed can neutralize the toxic RNA that causes two distinct repeat-expansion disorders, myotonic dystrophy 1 (DM1) and Fuchs endothelial corneal dystrophy (FECD). In the latter case, it can do so by an unexpected but powerful mechanism.

Genetic diseases in dire need of a treatment

DM1 is estimated to affect about 140,000 people in the United States. It can manifest anywhere from infancy to adulthood. And while it doesn’t always shorten lifespan, it often brings a debilitating set of symptoms including muscle weakness and pain, cataracts, and respiratory and gastrointestinal problems. The disorder is caused by a mutant copy of a gene called DMPK, whose RNAs contain dozens to hundreds of repeats of the RNA letters “CUG.”

FECD, which causes progressive damage to the cornea of the eye that often necessitates corneal transplantation, has a relatively high prevalence; studies suggest it manifests in at least several percent of Caucasian people older than 50. The disorder is caused by a mutant version of a gene called TCF4, whose RNAs also contain abnormally long CUG repeats.

These disorders arise from different mutant genes, and consequently appear in different cell types, but involve virtually the same toxic mechanism: In each case, the inclusion of an abnormally long sequence of CUG repeats causes the RNA copied from the gene to form structures that are “sticky” to certain other proteins in the cell, and effectively capture them—preventing them from doing their jobs in the cell. The depletion of one of these captured proteins, MBNL1, is a particularly important cause of cell damage and symptoms in DM1 and FECD.

Encouraging results in pre-clinical tests

For the new study, Disney and his team used advanced computational methods to design a small organic molecule that selectively binds to the abnormal CUG-expansion RNAs found in MD1- and FECD-affected cells, preventing these RNAs from capturing MBNL1.

To evaluate and improve the molecule, the team used a unique tool they had developed previously, Competitive Chem-CLIP, which allowed them to test their molecule’s ability to selectively recognize toxic CUG-expansion structures.

The team showed that in cultured cells derived from patients with DM1, as well as in an animal model of the disease, their improved designer molecule successfully reduced the depletion of MBNL1 and the loss of its function.

In FECD cells, the drug molecule also worked to prevent signs of disease, but this time by a different and potentially more powerful mechanism. In FECD cells, the disease-causing gene mutation occurs in a non-coding part of the gene called an intron. Normally, introns when copied into RNA are cut out of the RNA almost immediately and degraded by disposal systems in the cell. In FECD, the presence of the CUG-repeat expansion prevents the affected intron from being excised. However, Disney and his team found that their molecule allows that excision to take place, so that the abnormal RNA element is not just blocked but destroyed.

Targeting toxic RNAs with small organic molecules that can be put into pill form has generally been very challenging, so far, Disney notes, but the finding in this study points to the promising possibility of using such molecules not just to block bad RNAs but to trigger their destruction.

“If a drug causes a toxic RNA to be destroyed instead of merely blocking it, then the effect should be longer lasting,” he says.

Having performed their proof-of-principle demonstration, he and his team, which includes a startup biotech company, Expansion Therapeutics, are continuing to develop the molecule tested in the study as a potential drug treatment for DM1 and FECD.

The researchers also are taking a similar approach in developing potential drug treatments for RNA repeat-expansion diseases involving CAG repeats, which include the progressive and fatal neurological disorder known as Huntington’s disease.

Disney notes that his group’s computational approach to drug discovery, versus traditional methods involving the screening of large sets, or libraries, of molecules, gives them a big advantage: “Our ability to do computation-aided design allows us to get initial compounds quickly, and quickly test them,” Disney says.



More information:
Alicia J. Angelbello et al, A Small Molecule that Binds an RNA Repeat Expansion Stimulates Its Decay via the Exosome Complex, Cell Chemical Biology (2020). DOI: 10.1016/j.chembiol.2020.10.007