PermuteDDS

PermuteDDS predicts drug-drug synergy by integrating multiple representations of drugs and cancer cell lines using a permutable feature fusion network to prioritize synergistic drug combinations.

Key Features:

• Permutable Feature Fusion Network: Uses a permutable feature fusion network to integrate multiple representations of drugs and cell lines.
• Integration of Diverse Information Channels: Combines drug and cell line features across different channels to capture complex relationships relevant to synergy prediction.
• Performance on Benchmarks: Demonstrates performance on benchmark datasets and generalizes across independent test sets grouped by different tissues.
• Comparative and Ablation Experiments: Validated through comparative studies and ablation experiments to assess contribution of model components for drug-drug synergy prediction.

Scientific Applications:

• Identification of Synergistic Drug Combinations: Prioritizes drug pairs that exhibit potential synergistic effects for experimental follow-up.
• Oncology Combination Therapy Prioritization: Supports selection of combination therapies for cancer by evaluating drug interactions with cancer cell lines.
• Personalized Medicine and Tissue-Specific Analysis: Applies across different tissue-grouped test sets to inform tissue-specific and personalized treatment strategies.

Methodology:

Integrates multiple drug and cell line representations using a permutable feature fusion network and evaluates model performance with comparative and ablation experiments on benchmark datasets and independent tissue-grouped test sets.