iCDI-PseFpt

iCDI-PseFpt predicts ion channel–drug interactions from protein sequences and 2D drug fingerprints to support identification of drug targets such as ion channels involved in heartbeat, sensory transduction, and central nervous system responses.

Key Features:

• Sequence-based approach: Uses a sequence-based computational approach that circumvents the requirement for protein 3D structural data.
• Pseudo Amino Acid Composition (PseAAC): Represents protein ion-channel sequences using PseAAC implemented with gray model theory.
• 2D Molecular Fingerprint: Represents drug compounds by 2D molecular fingerprints to capture chemical properties without 3D structures.
• Fuzzy K-Nearest Neighbor: Employs a fuzzy K-Nearest Neighbor algorithm as the predictive classifier.
• Performance evaluation: Achieved an overall success rate of 87.27% under jackknife cross-validation, reported as superior to existing predictors.

Scientific Applications:

• Ion channel–drug interaction prediction: Predicts interactions between ion channels and small-molecule drugs relevant to heartbeat, sensory transduction, and central nervous system responses.
• Drug-target interaction networks: Applies the same sequence- and fingerprint-based approach to broader drug-target interaction network studies to inform drug development.

Methodology:

Pseudo Amino Acid Composition (PseAAC) using gray model theory; 2D molecular fingerprint representation of drugs; fuzzy K-Nearest Neighbor algorithm for prediction; jackknife cross-validation for performance evaluation.