RNA-Seq Analysis Strategies and Ethical Considerations Involved in Precision Medicine

Abstract

RNA-Seq has become the most recently and widely accepted method to evaluate gene expression. Though with RNA-Seq being a fairly green technology, analytical methods for its output data have not been fully investigated as they have for preceding technology; such as those methods used in analyses of microarray data. This is likely the result of the potential breadth of information that can be obtained from the different applications of RNA-Seq. Analyses of RNA-Seq data include: detecting differentially expressed genes, transcriptome profiling, and interpretation of gene functions. As with any advanced technology medical or otherwise, the longer it is available, the price of the technology, in general, decreases and the technology itself becomes more refined. This has been true for genomic sequencing—costs per sample have continued to decrease; and the accuracy and precision of results has improved greatly. Synchronously, more physicians have opted to have more of their patients’ genetic material sequenced. This has caused both challenges in the development of accurate, efficient, and consistent statistical methods; and much debate regarding the ethics involved in genomic sequencing. To provide insight into two statistical challenges that are common with analyzing RNA-Seq data, we conduct extensive simulation studies. These simulations studies include: 1) investigation of fitting complex models which account for pairedness across subject’s measurements in terms of the power gained and control of Type I error rate; and 2) evaluation of clustering performance of various clustering methods in transformed RNA-Seq data. In addition to investigating the aforementioned statistical challenges, we develop a protocol for a survey study which has the potential to provide insight into cancer patients’ opinions towards genomic sequencing as there is much ethics related controversy that surrounds the topic.