PyFeat

PyFeat generates sequence-derived features from DNA, RNA, and protein sequences to support prediction of structural, functional, interaction, and expression properties.

Key Features:

• Supported Sequence Types: Accepts DNA, RNA, and protein sequences as input for feature generation.
• Feature Extraction: Extracts a diverse set of features using 13 different techniques that capture interactions between neighboring residues.
• Efficient Feature Selection: Employs the AdaBoost algorithm to select features with maximum discriminatory potential and manage feature sparsity.
• Context-Free Combination: Represents context-free combinations of effective features derived from large neighboring residues for flexible feature set construction.

Scientific Applications:

• Protein Structure Prediction: Provides local residue interaction features to support prediction of protein structural properties.
• Functional Annotation: Supplies discriminatory sequence features useful for annotating protein and nucleic acid functions.
• Interaction Studies: Delivers feature sets that highlight potential molecular interaction sites within sequences.
• Expression Analysis: Enables analysis of expression-related patterns through extracted sequence-derived features.

Methodology:

Feature extraction using 13 different techniques that capture interactions between neighboring residues; AdaBoost for feature selection to reduce feature sparsity and retain maximally discriminatory features; representation of context-free combinations of effective features derived from large neighboring residues.