An artificial intelligence model developed by researchers at the Massachusetts Institute of Technology and the Swiss Federal Institute of Technology, Zurich (ETH Zurich) can forecast the binding configuration of two proteins.
The Equidock machine learning model targets rigid body docking, which transpires when two proteins bind by rotating or translating in three-dimensional (3D) space.
The model converts the proteins' 3D structures into 3D graphs that can be processed by the neural network; nodes in the graph represent individual amino acids, and geometric knowledge helps Equidock understand how objects can change when rotated or translated.
The model identifies protein atoms most likely to interact and form chemical bonds, then integrates the proteins into a complex.
Equidock could predict the final protein complex in one to five seconds, compared to four software models that required 10 to 60 minutes or more.
From MIT News
View Full Article
Abstracts Copyright © 2022 SmithBucklin, Washington, DC, USA
No entries found