PROFEAT

PROFEAT computes comprehensive structural, physicochemical, topological, interaction, atomic-level, and network descriptors from protein, peptide, and small-molecule sequences to support statistical learning models for predicting structural, functional, and interaction properties.

Key Features:

• Structural and Physicochemical Descriptors: Computes six feature groups composed of ten features, yielding 51 descriptors and 1447 descriptor values including amino acid composition, dipeptide composition, normalized Moreau-Broto autocorrelation, Moran autocorrelation, Geary autocorrelation, sequence-order-coupling number, quasi-sequence-order descriptors, and composition-transition-distribution of structural and physicochemical properties.
• Customizable Autocorrelations: Computes autocorrelation descriptors (Moreau-Broto, Moran, Geary) based on user-defined properties.
• Interaction Descriptors: Computes descriptors for protein-protein and protein-small molecule interactions.
• Segment Descriptors: Provides segment descriptors for analyzing local sequence properties.
• Topological and Atomic-level Descriptors: Includes topological descriptors for peptide sequences and small-molecule structures and atomic-level topological descriptors, with over 400 atomic-level topological descriptors for small molecules.
• Expanded Feature Groups: Includes pseudo-amino acid composition, amphiphilic pseudo-amino acid composition, and total amino acid properties.
• Network Descriptors: Computes up to 329 network and protein-protein interaction descriptors describing topology and connectivity of unweighted, edge-weighted, node-weighted, edge-node-weighted, and directed networks.

Scientific Applications:

• Predictive Modeling: Enables development of statistical learning models and quantitative structure-activity relationship (QSAR) models for predicting protein structural and functional classes, protein-protein interactions, peptide functions, and small-molecule activity.
• Systems-Level Investigations: Provides network descriptors for analysis of biological, disease, and pharmacological networks derived from genome, interactome, transcriptome, metabolome, and diseasome profiles.

Methodology:

Computes descriptors by calculating amino acid and dipeptide compositions; normalized Moreau-Broto, Moran, and Geary autocorrelations; sequence-order-coupling number; quasi-sequence-order descriptors; composition-transition-distribution; pseudo- and amphiphilic pseudo-amino acid compositions; topological and atomic-level descriptors; segment descriptors; and network topology descriptors.