Abstract:
|
The advent of Next-Generation Sequencing (NGS) technologies yield unprecedented in-depth sequencing information with offering vast opportunities in cancer genome research. Among NGS technologies, the RNA sequencing or RNA-seq technology has gained its popularity in transcriptomic profiling due to its finer resolution while yielding minimally biased signals. RNA-seq transcriptional profiling, however, can be easily confounded by normal tissue contaminations in tumor samples. To address this problem, copy number variation (CNV) from whole genome genotyping are often used for informing the purity of each tumor samples. Collecting both normal and tumor tissue RNA-seq data is expensive, thus reducing the cost of sampling tissues while maintaining power to detect differentially expressed (DE) genes is of interest in cancer research. Motivated by this, we like to propose a statistical tool that identifies the differentially expressed genes solely using tumor tissues that are not homogeneous as well as improve the accuracy in the tumor purity estimation compared the CNV only based tool.
|