SYNPRED
Predicting drug combination synergy across cancer cell lines using ensemble AI
0.91
F1-score
0.80
AUROC
0.85
Accuracy
0.91
Precision
SYNPRED is an ensemble AI framework that predicts the synergistic effects of drug combinations across cancer cell lines. Integrating omics and biophysical features with multiple synergy reference models, it provides both classification and regression outputs, with state-of-the-art performance on the NCI-ALMANAC dataset.
Available Features
- Single-pair prediction: Submit two SMILES files and a tissue type for instant synergy scoring
- Batch table upload: Screen multiple drug–cell line combinations in a single submission
- Five synergy metrics: Bliss, HSA, Loewe, ZIP, and CSS — choose the model most suited to your study
- Results by email: Results delivered to your inbox with a dedicated results page link
- Feature interpretability: Permutation importance and Sankey diagrams reveal which features drive predictions
- Stand-alone deployment: Download models and run SYNPRED locally via the GitHub repository
Ready to predict your next drug combination? Input two compounds, select your cancer cell line, and let SYNPRED do the rest – no programming knowledge required.