PIPR

PIPR (Protein-Protein Interaction Prediction Based on Siamese Residual RCNN) is an end-to-end deep learning framework for predicting protein-protein interactions (PPIs) using only protein sequences. Unlike traditional methods that rely on extracting predefined features from sequences, which can be costly and have limited coverage of PPI information, PIPR leverages a deep residual recurrent convolutional neural network (RCNN) in a Siamese architecture to capture both robust local features and contextualized information from the protein sequences.

The Siamese RCNN architecture in PIPR effectively captures the mutual influence between protein sequences, which is crucial for accurate PPI prediction. This deep learning approach eliminates the need for extensive data pre-processing and feature engineering, making it more generalizable to different application scenarios.

Experimental evaluations demonstrate that PIPR outperforms various state-of-the-art systems on the binary PPI prediction task. Moreover, it shows promising performance on more challenging problems, such as interaction type prediction and binding affinity estimation, where existing approaches often fall short.

Topic

Protein interactions;Machine learning

Detail

  • Operation: Residue contact prediction;Side chain modelling;Protein interaction prediction

  • Software interface: Command-line user interface

  • Language: Python, Shell

  • License: pache License 2.0

  • Cost: Free of charge with restrictions

  • Version name: -

  • Credit: The National Institutes of Health, the National Science Foundation.

  • Input: -

  • Output: -

  • Contact: Muhao Chen muhaochen@ucla.edu

  • Collection: -

  • Maturity: -

Publications

  • Multifaceted protein-protein interaction prediction based on Siamese residual RCNN.
  • Chen M, et al. Multifaceted protein-protein interaction prediction based on Siamese residual RCNN. Multifaceted protein-protein interaction prediction based on Siamese residual RCNN. 2019; 35:i305-i314. doi: 10.1093/bioinformatics/btz328
  • https://doi.org/10.1093/BIOINFORMATICS/BTZ328
  • PMID: 31510705
  • PMC: PMC6681469

Download and documentation


< Back to DB search