2020plus

2020plus is a computational tool to tackle the challenging task of distinguishing cancer driver genes from the vast number of somatic mutations identified in human cancers through sequencing technologies. As identifying driver genes is crucial for understanding the mechanisms of cancer development and for developing targeted therapies, the accuracy and reliability of such prediction tools are paramount.

2020plus sets itself apart by employing a machine-learning-based ratiometric approach to predict driver genes. This approach allows for a more nuanced analysis of somatic mutations, potentially improving the accuracy of driver gene identification. The tool is part of a broader evaluation framework that compares the performance of eight different driver gene prediction methods. This framework addresses the significant challenge of evaluating such methods without a gold standard—bona fide driver gene mutations.

The comprehensive analysis of this framework reveals that the driver genes predicted by each of the eight methods vary widely, highlighting the variability and the challenges in identifying true cancer drivers. Furthermore, inconsistencies in the reported P values among several methods raise questions about the underlying assumptions in these predictions, indicating areas where further methodological improvements could enhance prediction accuracy.