Identification of cancer driver genes and drug target genes by integrating multi-types of variation

Abstract

Cancer genome has a large number of alterations that occur in certain cells. These variations are caused by many factors, such as environmental factors and errors during cell division. However, not all of these alterations affect tumor development. Some alterations would affect cancer development and some would not. Therefore, it is difficult to classify them because alterations that do not affect cell development are more numerous than those that affect cell development. A previous study solved this problem by finding recurrent genes among large cohorts because the probability of recurrent genes is very low. As a result, recurrent genes play a key role in cancer development and we call them as driver genes. However, this conventional approach can only be applied to large cohorts and has limitations in analyzing an individual sample. Therefore, understanding an individual tumor requires a new approach. In this study, we built a pipeline that identifies driver genes for each sample using the whole genome sequence and whole transcriptome sequence data. In addition, we identified candidate driver genes that could be a drug target. Using this pipeline, we obtained driver genes and drug target genes for each sample. Our study shows that about five percentages of the alteration genes were found as driver genes and drug target genes in cancer.