protr

protr generates numerical representations of proteins and peptides from amino acid sequences as an R package for sequence-based descriptor computation and proteochemometric modeling.

Key Features:

• Numerical Representation Schemes: Provides a wide array of protein and peptide numerical descriptors derived from amino acid sequences.
• Customizable Descriptors: Supports selection and definition of amino acid properties from the AAindex database to create customized descriptors.
• Dimensionality Reduction: Implements six types of scale-based descriptors using multiple scaling methods for proteochemometric modeling applications.
• Similarity and Semantic Measures: Computes similarity scores from protein sequence alignment and Gene Ontology (GO) semantic similarity measures.
• Profile-Based Features: Calculates profile-based protein features using position-specific scoring matrices (PSSMs).

Scientific Applications:

• Structural Analysis: Provides numerical descriptors to elucidate structural properties of proteins and peptides.
• Functional Profiling: Enables functional exploration using GO semantic similarity measures and profile-based features.
• Proteochemometric Modeling: Supports proteomics and cheminformatics modeling through scale-based descriptors and dimensionality reduction.

Methodology:

Computes numerical descriptors from amino acid sequences, uses AAindex-derived amino acid properties, derives six types of scale-based descriptors, computes sequence alignment-based similarity scores and Gene Ontology semantic similarity, and extracts profile-based features from position-specific scoring matrices (PSSMs).