Freescore

Freescore computes alignment-based scores to extract features from conotoxin protein sequences for support vector machine (SVM) classification, facilitating sequence-based analysis relevant to drug design.

Key Features:

• Local alignment partition functions: Uses local alignment partition functions to score sequence alignments for feature extraction.
• Suboptimal alignments of restricted-length words: Considers suboptimal alignments of words with restricted lengths to capture near-optimal sequence matches.
• Feature extraction for SVM: Generates alignment-derived features specifically for classification with support vector machines (SVM), implemented as SVM-Freescore.
• Performance improvement metrics: Reports an approximate sensitivity increase of 5.864% and specificity increase of 3.76% relative to previously reported methods.
• High computed metrics: Achieves average computed sensitivity of 0.9742 and specificity of 0.9917 on benchmark datasets.
• Sequence-based analysis: Enables sequence-based analysis for functional and structural characterization of conotoxin proteins.
• Database application: Applied to superfamily classification within curated databases such as ConoServer.

Scientific Applications:

• Conotoxin protein classification: Classifies conotoxin proteins using alignment-derived features and SVM-Freescore.
• Superfamily classification: Performs superfamily-level classification of conotoxins in curated databases such as ConoServer.
• Functional and structural characterization: Supports sequence-based inference of conotoxin peptide function and structural relationships.
• Drug design research support: Provides classification and feature information relevant to drug design targeting schizophrenia, neuromuscular conditions, and chronic pain.

Methodology:

Uses local alignment partition functions that consider suboptimal alignments of words with restricted lengths to extract features, followed by support vector machine classification (SVM-Freescore).