Evaluating Protein, Binding Interface with Transformer Networks


Principal Investigator: Prof. Giri Narasimhan
Principal Architect: Vitalii Stebliankin
Other Contributors: Azam Shirali, Prabin Baral, Jimeng Shi, Prem Chapagain, Kalai Mathee



PIsToN (Evaluating Protein, Binding Interface with Transformer Networks) - is a novel deep learning-based approach for distinguishing native-like protein complexes from decoys. Each protein interface is transformed into a collection of 2D images (interface maps), where each image corresponds to a geometric or biochemical property in which pixel intensity represents the feature values. Such data representation provides atomic-level resolution of relevant protein characteristics. To build hybrid machine learning models, additional empirical-based energy terms are computed and provided as inputs to the neural network. The model is trained on thousands of native and computationally predicted protein complexes that contain challenging examples. The multi-attention transformer network is also endowed with explainability by highlighting the specific features and binding sites that were the most important for the classification decision.
Download Github Site: PIsToN
Supplementary notebook Python code
Contact: Prof. Giri Narasimhan



Vitalii Stebliankin, Azam Shirali, Prabin Baral, Jimeng Shi, Prem Chapagain, Kalai Mathee, and Giri Narasimhan. Evaluating Protein Binding Interfaces with Transformer Networks, Nature Machine Intelligence, Sep 2023.


Vitalii Stebliankin