PaccMann

PaccMann predicts drug sensitivity by integrating transcriptomic cell line profiles, compound structure information, and drug sensitivity screening data for molecularly informed anticancer compound prioritization.

Key Features:

• Multimodal Data Fusion: Integrates transcriptomic cell line profiles, compound structural information, and drug sensitivity screening data to inform predictions.
• Neural Network Architecture: Employs advanced neural networks for drug sensitivity prediction across integrated data modalities.
• In-Silico Testing: Enables virtual evaluation of transcriptomic profiles against compound structures for predictive assessment.
• Attention-based Interpretability: Uses attention-based neural networks that focus on relevant genes and chemical substructures and provides confidence scores for interpretation of predictions.

Scientific Applications:

• Drug Repositioning: Predicts potential new therapeutic uses of existing drugs for specific cancer biomolecular samples.
• Lead Compound Identification: Assists in prioritizing promising lead compounds using predictive analysis of transcriptomic and chemical data.

Methodology:

Integrates transcriptomic cell line profiles, compound structural representations, and drug sensitivity screening data using attention-based neural networks that highlight relevant genes and chemical substructures and output confidence scores for interpretability.