CORNET
CORNET predicts residue contacts within proteins using neural networks and chemo-physical plus evolutionary data to generate contact maps for analysis of protein structure and function.
Key Features:
- Neural network architecture: Employs neural networks trained on 200 non-homologous proteins with well-resolved three-dimensional structures using standard backpropagation to learn associations between covalent protein structure and contact maps.
- Input data: Integrates chemo-physical properties and evolutionary information as inputs for contact prediction.
- Validation dataset: Validated against an independent test set of 408 proteins with known structures that are non-homologous to the training set.
- Performance comparison: Demonstrates improved accuracy relative to previous statistical approaches for predicting protein contact maps.
Scientific Applications:
- Protein structure prediction: Provides contact maps that aid in modeling protein tertiary structure from sequence-derived information.
- Structural biology research: Enables inference of structural features from sequence and derived properties for proteins lacking experimental structures.
- Drug design and development: Supplies residue contact information useful for identifying interaction sites and informing ligand design.
Methodology:
Train neural networks with standard backpropagation using chemo-physical properties and evolutionary information as inputs to learn associations between covalent protein structure and contact maps; training used 200 non-homologous proteins with well-resolved three-dimensional structures and validation used an independent test set of 408 non-homologous proteins.
Topics
Collections
Details
- Maturity:
- Mature
- Cost:
- Free of charge
- Tool Type:
- web application
- Operating Systems:
- Linux, Windows, Mac
- Added:
- 1/22/2015
- Last Updated:
- 11/24/2024
Operations
Data Inputs & Outputs
Protein secondary structure prediction
Publications
Fariselli P, Casadio R. A neural network based predictor of residue contacts in proteins. Protein Engineering, Design and Selection. 1999;12(1):15-21. doi:10.1093/protein/12.1.15. PMID:10065706.