Use of Synthetic Transcript Pools to Evaluate RNA-Seq Performance and Analytical Methods
Year: 2011
Type: article
Abstract: RNA-Seq and related Massively Parallel Sequencing methods are rapidly becoming standardized platforms for transcriptome studies. Although RNA-Seq is now being used routinely, there has not been a thorough investigation of fundamental performance metrics using standardized reagents and analytics. In order to evaluate the capabilities of RNASeq and associated analysis methods we have used two synthetic RNA spike-in pools that can be added to the RNA sample prior to library preparation and sequencing. The pools consist of 92 synthetic transcripts, designed by the External RNA Control Consortium (ERCC). The control pools are designed to enable assessment of both within and between sample performance. Using sequencing data generated with these controls we have benchmarked several statistical methods designed specifically to assess differential expression of count data. We show that methods which rely on Poisson distributions dramatically underestimate the variance associated with RNA-Seq data. This underestimation causes an inflated False Positive Rate. We also show that various methods for estimating the dispersion parameter in Negative Bionomial tests also vary in their ability to control the Type I error rate. These performance differences are true in the absence of biological variability, suggesting that even technical variability associated with sequencing is not adequately described by a Poisson distribution. Additionally, we show that the controls can be used to evaluate performance of various normalization and scaling methods for RNA-Seq count data. This analysis suggests that traditional scaling methods based on the total number of reads result in a significant loss of accuracy and more robust methods are often required. This analysis reveals the importance of using standardized reagents and controls as a routine part of the RNA sequencing workflow.
Source: Europe PMC (PubMed Central)
Authors: Qu L, Andrew L. Lemire, Joel Brockman, Sheila J. Heater, Jeff Schageman, Jiang Gu, Lea K, San Jose C, Natalie Hernandez, D. Batten, K. Bramlett, Diane Ilsley, Christopher G. Mueller, Robert A. Setterquist, Penn Whitley
Language: English
Subfield: Molecular Biology
Domain: Life Sciences
Open Access status: green
OpenAlex ID: W2328447830