Detecting cancer-driving mutations is crucial for understanding cancer and developing treatments. Existing prediction tools vary in accuracy. N2CR member A/Prof Chen Ee Sin co-led the study on DriverDetect, a machine learning algorithm (developed in NUS) that combines outputs from seven tools to better predict driver mutations. Trained on cancer-specific mutation datasets, it outperformed individual tools in validation tests. DriverDetect can integrate new prediction algorithms and be retrained with new data, making it suitable for broad cancer analysis and cross-cancer studies.

Click here to read more.