Using mixtures of biological samples as genome-scale process controls

Published

September 17, 2015

Author(s)

Jerod R. Parsons, Patrick S. Pine, Marc L. Salit, Sarah A. Munro, Jennifer H. McDaniel

Abstract

Background: Genome-scale “-omics” measurements are challenging to benchmark, due to the enormous number of individual measurands. Mixtures of previously-characterized samples can be used to benchmark repeat performance using mixture proportions as truth for the measurement. We describe and evaluate experiments characterizing the performance of RNA-sequencing (RNA-Seq) measurements. Results: The parameters of a model fit to a measured -omic profile can be evaluated to assess bias and variability of the genome-scale measurement of a mixture. A linear model describes the behavior of expression measures of mixtures and provides a context for performance benchmarking. Experimental residuals from the model can be used as a metric to evaluate the magnitude of effect the experimental process has on the linearity, additivity, and precision of the underlying measurement. Accurate benchmarking requires that the mixture proportions be well-defined, which for RNA-Seq requires knowledge of the messenger RNA (mRNA) content of the mixture components. We demonstrate and evaluate an experimental method suitable for use in a genome-scale process control and lay out a method utilizing spike-in controls to determine mRNA content. Conclusions: Genome-scale process controls can be derived from mixtures. These controls measure the repeatability and additive linearity of measurements by relating prior knowledge of the mixture composition components to measurements in a complex mixture. The mRNA ratio between mixture components becomes an important term to experimentally derive their proportions in a transcriptomic mixture. Spike-in controls can be utilized to measure the relationship between mRNA content and input total RNA. Even when gene expression markers are not previously known, deconvolution using this method can return the proportions of known components in an unknown mixture.

Citation

BMC Genomics

Volume

16

Pub Type

Journals

Download Paper

DOI Link

Keywords

RNA sequencing, Gene expression, mixture deconvolution, expression deconvolution, process control, spike-in control, ERCC

Citation

Parsons, J. , Pine, P. , Salit, M. , Munro, S. and McDaniel, J. (2015), Using mixtures of biological samples as genome-scale process controls, BMC Genomics, [online], https://doi.org/10.1186/s12864-015-1912-7 (Accessed October 14, 2025)