Welcome to PLM-interact’s documentation!

PLM-interact is a Python library for PPI prediction and mutant effects prediction.

Protein structure can now be predicted from amino acid sequences with unprecedented accuracy, yet the prediction of protein–protein interactions remains a challenge. Here, we demonstrate that protein language models originally designed for single proteins can be extended to jointly encode interacting protein pairs. Our model, PLM-Interact, achieves state-of-the-art performance on both within-species and cross-species PPI prediction benchmarks and can also be fine-tuned to predict mutation effects in human PPIs.

Check out the Usage section and Installation section for how to install and use this tool for PPI prediction.

Note

This project is under active development.

Contents

• Installation
  • Install with pip
  • Install from sources
• Usage
  • Quick start
    • There are 6 commands in PLM-interact package
    • Predict a list of PPIs
    • Training PPI models
    • Validation and test of trained models
    • Fine-tuning in the binary mutation effect task.
    • Inference in the binary mutation effect task.
  • PPI prediction
  • Training PPI models using mask and binary classification losses
  • Training PPI models using only binary classification loss.
  • Evaluation and test with multi-nodes and multi-GPUs
  • Fine-tuning in the binary mutation effect task.
  • Inference in the binary mutation effect task.
• Data
  • Trained Models
  • Datasets for PLM-interact
  • Original Datasets
  • References