|
Abstract:
|
Single-subject RNA-Seq analysis unveils personalized mechanisms in disease progression and, therefore, holds great promise for precision medicine. Identifying personalized sets of the differentially expressed genes (DEG) is key in discovering personalized disease mechanisms. By virtue of next-generation sequencing, tens of thousands of genes (high dimension, p) of a single subject can be evaluated simultaneously for differential expression via RNA-Sequencing (RNA-Seq). On the other hand, many studies have no replicates (n = 1) for any individual, due to the intention to avoid heterogeneity from multiple samples, the limited accessibility of scarce tissue samples, or the significant cost of the high-through technology. A problem with this nature, large p and n=1, presents a high challenge to statistical methods. To address this challenge, we propose a novel method, which identifies differentially expressed genes from only two transcriptomes (e.g. normal vs. cancerous) of an individual. We compare operating characteristics of our method with other available methods in a simulation study and illustrate the new method with an application to a breast cancer dataset.
|
