PPAI

PPAI predicts aptamers and protein-aptamer interactions using machine learning to support identification and analysis of binding partners.

Key Features:

• Machine Learning Framework: Integrates Adaboost and Random Forest algorithms for prediction of aptamers and protein-aptamer interactions.
• Protein Feature Extraction: Extracts amino acid composition, pseudo-amino acid composition, grouped amino acid composition, C/T/D composition, and sequence-order-coupling number from protein sequences.
• Aptamer Feature Extraction: Utilizes nucleotide composition, pseudo-nucleotide composition (PseKNC), and normalized Moreau-Broto autocorrelation coefficient for aptamer sequences.
• Data Balancing: Applies the SMOTE algorithm to balance samples and address class imbalance.
• Performance Validation: Validated on an independent test set with AUCs of 0.907 for aptamer prediction and 0.871 for protein-aptamer interaction prediction.

Scientific Applications:

• Protein-Aptamer Interaction Analysis: Enables prediction-driven exploration of the biological roles of protein-aptamer interactions.
• Aptamer Discovery and Therapeutic Development: Supports identification of candidate aptamers to facilitate development of aptamer-based therapeutic strategies.

Methodology:

Extracts specified protein features (amino acid composition, pseudo-amino acid composition, grouped amino acid composition, C/T/D composition, sequence-order-coupling number) and aptamer features (nucleotide composition, PseKNC, normalized Moreau-Broto autocorrelation coefficient), applies SMOTE for class balancing, and uses Adaboost and Random Forest algorithms with validation on an independent test set (AUCs: 0.907 aptamer prediction; 0.871 protein-aptamer interaction prediction).