How to normalize metatranscriptomic count data for differential expression analysis

Differential expression analysis on the basis of RNA-Seq count data has become a standard tool in transcriptomics. Several studies have shown that prior normalization of the data is crucial for a reliable detection of transcriptional differences. Until now it has not been clear whether and how the transcriptomic approach can be used for differential expression analysis in metatranscriptomics.

University of Goettingen researchers propose a model for differential expression in metatranscriptomics that explicitly accounts for variations in the taxonomic composition of transcripts across different samples. As a main consequence the correct normalization of metatranscriptomic count data under this model requires the taxonomic separation of the data into organism-specific bins. Then the taxon-specific scaling of organism profiles yields a valid normalization and allows us to recombine the scaled profiles into a metatranscriptomic count matrix. This matrix can then be analyzed with statistical tools for transcriptomic count data. For taxon-specific scaling and recombination of scaled counts they provide a simple R script.

When applying transcriptomic tools for differential expression analysis directly to metatranscriptomic data with an organism-independent (global) scaling of counts the resulting differences may be difficult to interpret. The differences may correspond to changing functional profiles of the contributing organisms but may also result from a variation of taxonomic abundances. Taxon-specific scaling eliminates this variation and therefore the resulting differences actually reflect a different behavior of organisms under changing conditions. In simulation studies we show that the divergence between results from global and taxon-specific scaling can be drastic. In particular, the variation of organism abundances can imply a considerable increase of significant differences with global scaling. Also, on real metatranscriptomic data, the predictions from taxon-specific and global scaling can differ widely. These studies indicate that in real data applications performed with global scaling it might be impossible to distinguish between differential expression in terms of transcriptomic changes and differential composition in terms of changing taxonomic proportions.

Work flow for taxon-specific normalization


(A) Sequence samples from conditions A (white) and B (light gray). Assign each sequence read to taxonomic and feature categories. (B) Compute feature profiles from the assignment counts. (C) Obtain count matrix from taxon-specific feature profiles. (D) Normalize feature profiles of each taxon-specific count matrix separately. (E) Recombine normalized feature profiles of all taxa into a metatranscriptomic profile. (F) Perform differential expression analysis on metatranscriptomic count matrix.

As in transcriptomics, a proper normalization of count data is also essential for differential expression analysis in metatranscriptomics. This model implies a taxon-specific scaling of counts for normalization of the data. The application of taxon-specific scaling consequently removes taxonomic composition variations from functional profiles and therefore provides a clear interpretation of the observed functional differences.

Availability -R-script to normalize and analyze metatranscriptomic count data with taxon-specific scaling and DESeq2:

Klingenberg H, Meinicke P. (2017) How to normalize metatranscriptomic count data for differential expression analysis. PeerJ 5:e3859. [article]

Leave a Reply

Your email address will not be published. Required fields are marked *


Time limit is exhausted. Please reload CAPTCHA.