Giannoukos and colleagues benchmarked rRNA-depletion methods for bacterial and complex-community RNA-seq. Their model combined Prochlorococcus marinus, Escherichia coli, and Rhodobacter sphaeroides, spanning approximately 30%, 51%, and 69% genomic GC content.
The study compared five depletion approaches, including light normalization with duplex-specific nuclease. Residual rRNA and coding-sequence reads were measured by sequencing, and the investigators also assessed how well expression values were preserved after depletion.

Published Figure 1: Performance evaluation of five rRNA depletion methods. Reproduced from the Genome Biology source article under CC BY 2.0 with attribution.
For DSN light normalization, rRNA reads were reduced from approximately 98% to 10% for P. marinus and from approximately 98% to 7% for E. coli. In contrast, the GC-rich R. sphaeroides library changed from 99.5% to 76% rRNA after depletion, and the reported coefficient of determination for expression values was R2 = 0.13.
This independent study illustrates why DSN-based depletion should be evaluated as a sample-dependent strategy rather than marketed as universally species-independent. We use species composition, GC context, analytical objective, and post-depletion QC to decide whether a DSN route is appropriate. This is a published literature example, not a CD Genomics customer project or performance guarantee.


Figure 1. Read composition before and after depletion
Figure 2. Expression-profile concordance
Figure 3. Informative transcript representation