FreshAirTonight/af2complex
Predicting direct protein-protein interactions with AlphaFold deep learning neural network models.
Added some experimental options and improvements. For example, an option to pair MSA by top N per species or organism.
Support AF-Multimer v3 multimer models released by DeepMind in AlphaFold version 2.3.0 More options for input feature generation
📋 Changes
- Google Colab notebook access
- Added multiple MSA pairing options
- Enhanced interface score assessment
- Clustering predicted complexes
Modified the example to demonstrate an application of the new AF-Multimer version 2 models.
Minor bug fixes and code refactoring.
Updated version supports both monomer and multimer neural network models of AlphaFold for predicting protein complex.
First public release of AF2Complex. Predicting direct protein-protein interactions in multimeric proteins using AlphaFold neural network models. Developed and tested with AlphaFold v2.0.1.
