dipcell

dipcell predicts inhibitory activity of unknown molecules and facilitates design of analogs targeting pancreatic cancer cell lines.

Key Features:

• QSAR models: Quantitative Structure-Activity Relationship models trained on extensive pharmacological data predict inhibitory activity and promiscuous inhibitors for pancreatic cancer cell lines.
• Model performance: Models achieved a maximum Pearson correlation coefficient of 0.86 under 10-fold cross-validation.
• Drug-to-oncogene relationships: Integrated models validate drug-to-oncogene relationships linking compounds to specific oncogenes relevant in pancreatic cancer.
• FDA-approved drug screening: Computational screening of FDA-approved drugs identifies candidates for repurposing or modification against pancreatic cancer.
• Effective and resistant drug identification: The approach identifies the most and least effective drugs across pancreatic cancer cell lines and flags resistant cell lines.
• In vitro validation: Identified candidate drugs have been tested in vitro against pancreatic cancer cell lines.

Scientific Applications:

• Novel drug screening: Prioritizes novel promiscuous drug molecules with potential inhibitory effects on pancreatic cancer.
• Analog design: Guides design of analogs of known compounds to enhance efficacy or mitigate resistance.
• Resistance investigation: Identifies resistant pancreatic cancer cell lines to inform studies of resistance mechanisms.

Methodology:

Quantitative Structure-Activity Relationship (QSAR) modeling trained on pharmacological data, evaluated by 10-fold cross-validation (maximum Pearson correlation coefficient = 0.86), with computational screening of FDA-approved drugs and validation of drug-to-oncogene relationships.