Researchers at the University of Chicago have developed machine learning tools that can screen the best candidates of self-assembled peptides in an effort to develop electronic devices that are compatible with the human body.
The system used active learning or Bayesian optimization to guide molecular simulations, and to construct reliable data-driven models of how the sequence of the peptide influenced its properties after considering just 186 peptides.
The model predictions, which the researchers said were devoid of human bias, could then be reliably extrapolated to predict the properties of the rest of the peptide family.
The team ranked each peptide and gave the results to experimental collaborators, who will test the top candidates in the lab.
The researchers hope to use the machine learning system to design proteins, optimize self-assembling colloids to make atomic crystals, and to incorporate these tools into an autonomous laboratory.
From University of Chicago
View Full Article
Abstracts Copyright © 2020 SmithBucklin, Washington, DC, USA