google-deepmind/alphafold
Alphafold
AlphaFold is a deep learning biology tool and structural bioinformatic pipeline designed to predict the three-dimensional shapes of proteins from their amino acid sequences. It functions as a machine learning system capable of generating 3D molecular models for both monomeric proteins and multimeric protein complexes, including homomers and heteromers.
The system incorporates evolutionary information through multiple sequence alignment to identify physical proximity between residues. It utilizes a neural network architecture featuring spatial attention mechanisms and iterative refinement to determine the final atomic coordinates of a protein structure.
The pipeline includes capabilities for protein structural refinement via a relaxation procedure to resolve stereochemical violations. It also provides confidence analysis, generating scores to identify high-confidence sequence regions and evaluate global domain packing.
Features
- Protein Structure Prediction - Computes the three-dimensional shape of a protein based on its amino acid sequence using deep learning.
- Equivariant - Implements invariant point attention to process 3D coordinates regardless of the protein's rotation or shift.
- Geometry Confidence Metrics - Calculates per-residue and global error estimates to evaluate the local consistency and reliability of the predicted 3D geometry.
- Refinement Modules - Provides a module that recursively refines 3D atomic coordinates through iterative neural network passes.
- Protein Representation Networks - Uses an Evoformer-based transformer architecture to iteratively refine protein sequence and pair representations.
- Structural Confidence Analysis - Generates pLDDT and predicted TM-scores to identify high-confidence sequence regions and assess global domain packing.
- Protein Folding Inference - Computes the 3D structure of a protein from its primary amino acid sequence using a machine learning inference pipeline.
- Structure Prediction Models - Functions as a machine learning system that predicts the 3D shape of protein complexes and monomers from amino acid sequences.
- Multimeric Structure Prediction - Predicts the structure of multimeric proteins, including homomers and heteromers, by processing multiple amino acid sequences.
- Evolutionary Sequence Integration - Incorporates evolutionary information via multiple sequence alignment to identify co-varying residues and physical proximity.
- Deep Learning Biology Frameworks - Provides a deep learning computational framework designed to solve the protein folding problem for structural biology research.
- Protein Complex Prediction - Determines the spatial arrangement of multimeric proteins, including homomers and heteromers.
- Protein Interaction Modeling - Predicts the 3D spatial arrangement and interaction of multimeric protein complexes, including homomers and heteromers.
- Protein Reliability Analysis - Evaluates the reliability of predicted protein structures using scoring metrics to identify high-confidence regions.
- Structural Bioinformatics Analysis - Provides a set of inference tools that generate 3D molecular models and confidence scores for protein folding analysis.
- Stereochemical Refinement - Performs an Amber relaxation procedure on unrelaxed model outputs to resolve stereochemical violations in the final structure.
- Molecular Dynamics Force Fields - Implements an Amber relaxation procedure to resolve stereochemical violations and improve the physical stability of predicted structures.
- Protein Structural Refinement - Applies relaxation procedures to predicted models to fix stereochemical errors and improve the accuracy of the final structure.
Star history
How this analysis was created: This summary and feature list were written by an AI model that read the project's README and public documentation pages. Each feature links to the documentation it came from; stars, license and language come straight from the GitHub API. The model does not read the source code, and the analysis is refreshed when the project is re-analysed. Learn more on our About page.
