Power analysis for RNA sequencing and mass spectrometry-based proteomics data

Abstract

RNA-sequencing and mass spectrometry technologies have facilitated the differential expression discoveries in transcriptome and proteome studies. However, the determination of sample size to achieve adequate statistical power has been a major challenge in experimental design. The objective of this study is to develop a power analysis tool applicable to both RNA-seq and MS-based proteomics data. The methods proposed in this study are capable of both prospective and retrospective power analyses. In terms of the performance, the benchmarking results indicated that the proposed methods can give distinct power estimates for both differentially and equivalently expressed genes or proteins without prior differential expression analysis and other assumptions, such as, expected fraction of differentially expressed features, minimal fold changes and expected mean expressions. Using the proposed methods, not only can researchers evaluate the reliability of their acquired significant results, but also estimate the sufficient sample size for a desired power. The proposed methods in this study were implemented as an R package, which can be freely accessed from Bioconductor project at http://bioconductor.org/packages/PowerExplorer/.