ICTC-RAAC

ICTC-RAAC predicts ion channel-targeted conotoxins by applying the Reduced Amino Acid Cluster (RAAC) methodology to identify functionally conserved regions in conotoxin sequences for analysis of ion channel interactions.

Key Features:

• RAAC generation: Produces 673 RAACs derived from 74 reduced amino acid alphabets to represent reduced-sequence complexity.
• Variance-maximization algorithm: Uses an amino-acid reduction of variance maximization algorithm to distinguish conotoxin types.
• Tripeptide composition framework: Employs a tripeptide composition scheme with N=3 for sequence representation.
• Performance configuration: Achieved 89.3% accuracy using the Type 20, Cluster 9 configuration within the tripeptide composition framework.
• Feature selection: Applies ANOVA with incremental feature selection (IFS) to refine predictive feature sets.
• Validation: Uses cross-validation reporting a best overall accuracy of 96.4%, exceeding prior results by 1.8%.
• Conotoxin-specific handling: Accounts for disulfide-rich and sequence-diverse conotoxins and identifies functionally conserved regions.

Scientific Applications:

• Ion channel-targeted conotoxin identification: Predicts conotoxins that target ion channels based on sequence-derived RAAC features.
• Conotoxin type classification: Distinguishes conotoxin types using variance-maximization-derived features and tripeptide composition.
• Drug design and neuropharmacology prioritization: Supports selection of conotoxin candidates for pharmacological development and study of ion channel interactions.

Methodology:

Generates 673 RAACs from 74 reduced amino acid alphabets, represents sequences using tripeptide composition (N=3), applies an amino-acid reduction of variance maximization algorithm, refines features with ANOVA plus incremental feature selection (IFS), and evaluates performance by cross-validation.