newDNA-Prot

newDNA-Prot classifies proteins as DNA-binding or non-DNA-binding using an SVM trained on a comprehensive set of sequence-derived and functional features to support studies of gene regulation.

Key Features:

• Support Vector Machine Classifier: Utilizes a support vector machine (SVM) classifier to distinguish DNA-binding from non-DNA-binding proteins.
• Comprehensive Feature Representation: Employs an extensive feature set grouped into six categories: Primary Sequence Based, Evolutionary Profile Based, Predicted Secondary Structure Based, Predicted Relative Solvent Accessibility Based, Physicochemical Property Based, and Biological Function Based.
• Feature Selection Methods: Applies mRMR (Minimum Redundancy Maximum Relevance), a Wrapper method, and a two-stage feature selection method, with the two-stage method reported to outperform the others in removing irrelevant features and reducing redundancy.
• Statistical Analysis: Performs statistical analysis of selected features, with over 95% of features reported as statistically significant across the six feature groups.

Scientific Applications:

• Gene regulation studies: Predicts DNA-binding proteins to support research into gene regulation mechanisms.
• Genomics and molecular biology prediction and benchmarking: Provides predictive performance that outperforms methods such as iDNA-Prot, DNAbinder, and DNA-Prot, with reported improvement of approximately 10% over DNA-Prot across various evaluation measures.

Methodology:

Uses an SVM classifier trained on features from the six specified groups with feature selection by mRMR, Wrapper, and a two-stage method, followed by statistical analysis of selected features (>95% significant).