Meta-analysis of genomic and proteomic features to predict synthetic lethality of yeast and human cancer

Abstract

A major goal in cancer medicine is to find selective drugs with reduced side-effect. A pair of genes is called synthetic lethality (SL) if mutations of both genes will kill a cell while mutation of either gene alone will not. Hence, a gene in SL interactions with a cancer-specific mutated gene will be a promising drug target with anti-cancer selectivity. Wet-lab screening approach is still so costly that even for yeast only a small fraction of gene pairs has been covered. Computational methods are therefore important for large-scale discovery of SL interactions. Most existing approaches focus on individual features or machine learning methods, which are prone to noise or overfitting. In this paper, we propose an approach of meta-analysis that integrates 17 genomic and proteomic features and the outputs of 10 classification methods. It thus combines the strengths of existing methods. It also adjusts relative contributions of multiple methods with weights learned from the training data. Running on a dataset of the yeast strain of S. cerevisiae, our method achieves AUC (area under ROC curve) of 87.2% the highest among all competitors. Moreover, through orthologous mapping from yeast to human genes, we predicted a list of SL pairs in human that contain top mutated genes in lung and breast cancers recently reported by The Cancer Genome Atlas (TCGA). Our method and predictions would shed light on mechanisms of SL and lead to discovery of novel anti-cancer drugs.