Skip to main page content
Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
, 14 (10), e0224578
eCollection

Evaluation of Protocols for rRNA Depletion-Based RNA Sequencing of Nanogram Inputs of Mammalian Total RNA

Affiliations

Evaluation of Protocols for rRNA Depletion-Based RNA Sequencing of Nanogram Inputs of Mammalian Total RNA

Simon Haile et al. PLoS One.

Abstract

Next generation RNA-sequencing (RNA-seq) is a flexible approach that can be applied to a range of applications including global quantification of transcript expression, the characterization of RNA structure such as splicing patterns and profiling of expressed mutations. Many RNA-seq protocols require up to microgram levels of total RNA input amounts to generate high quality data, and thus remain impractical for the limited starting material amounts typically obtained from rare cell populations, such as those from early developmental stages or from laser micro-dissected clinical samples. Here, we present an assessment of the contemporary ribosomal RNA depletion-based protocols, and identify those that are suitable for inputs as low as 1-10 ng of intact total RNA and 100-500 ng of partially degraded RNA from formalin-fixed paraffin-embedded tissues.

Conflict of interest statement

The authors have declared that no competing interests exist.

Figures

Fig 1
Fig 1. rRNA and mitochondrial transcript content.
Ribo-zero Gold (RZG) vs. NEB RNase H-based rRNA depletion protocol (RNase H). Input was UHR total RNA at the indicated total RNA input amounts. (A) Reads aligning to 18S and 28S rRNA. (B) Reads aligning to 45S rRNA. (C) Mitochondrial RNA content. (D) Relative levels of each of the mitochondrial mRNAs between the two protocols as compared to the levels of mitochondrial rRNAs.
Fig 2
Fig 2. Diversity, regional mapping and expression correlations.
Ribo-zero Gold (RZG) vs NEB RNase H-based rRNA depletion protocol (RNase H). Input was UHR total RNA at the indicated total RNA input amounts. (A) Proportions of duplicate reads. (B) Proportions of exonic, intronic and intergenic reads. (C) Expression correlations across RNA input amounts. Pearson’s correlation coefficient was calculated pair-wise for all transcripts.
Fig 3
Fig 3. Validation of expression quantification accuracy.
Ribo-zero Gold (RZG) vs NEB RNase H-based rRNA depletion protocol (RNase H). Input was UHR total RNA at the indicated total RNA input amounts. (A) qPCR data for ~1,000 mRNAs [23] compared to RNA-seq data. (B) Correlation of observed versus expected ERCC spike-in levels. (C) Log-log plots of observed versus expected ERCC RNAs. Blue dots represent amounts of individual spike-in RNAs, the number of which is variable between libraries depending on the detection sensitivity of the protocol.
Fig 4
Fig 4. rRNA content comparisons using FFPE samples.
Ribo-zero Gold (RZG) vs NEB RNase H-based rRNA depletion protocol (RNase H). Input was FFPE total RNA at the indicated total RNA input amounts. The two samples, FFPE-1 and FFPE-2, are described in the text. Reads mapping to the 3’-external spacer (3’ES) are shown in the left panels (top and bottom) and reads mapping to other regions of the 45S precursor RNA are shown in the middle panels (top and bottom). RNA size profiles from Agilent RNA Nano assays are shown in the right panels. In red are the profiles for the RNA input before DNase I treatment (Input) and in blue are profiles for RNA after DNase I treatment (Post-DNase). Vertical arrows delineate indicated sizes in nucleotides (nt) and the proportions of fragments between 200 and 5000 nt are indicated in the insets.
Fig 5
Fig 5. Expression correlation and hierarchical clustering of data from matched fresh-frozen and FFPE- derived samples (n = 39).
(A) Pearson’s correlation of transcript levels between fresh-frozen and FFPE samples (Y-axis) for various total RNA input amounts (X-axis). (B) Hierarchical clustering. Variance-stabilized expression values for 1,000 genes whose expression was most variable were chosen for clustering. Samples were hierarchically clustered based on inter-sample Pearson correlation values. The results indicate that FFPE preparation of samples does not result in a dominant batch effect that occludes the biological source of the material (i.e., the patient’s tumour).
Fig 6
Fig 6. Analysis of fusion transcripts in the RNase H protocol.
25 ng and 100 ng UHR total RNA input libraries were evaluated for the detection of events that were previously validated using qPCR [33]. Filled gray boxes indicate that events were positively identified. qPCR cycle threshold (Ct) data are from [33].
Fig 7
Fig 7. Effects of lowering input amounts using the RNase H protocol.
The input was UHR total RNA at indicated total RNA input amounts. (A) Proportion of aligned reads achieved using 1–100 ng of total RNA input. (B) Genes identified as a function of input amount. Blue indicates genes identified with greater than 0 RPKM values; red indicates genes identified with greater than 1 RPKM values, and green indicate genes identified with greater than 5 RPKM values (indicative of more abundant transcripts). (C) Expression correlations across UHR RNA input amounts, indicated by numbers on both axes. Pearson’s correlation coefficient was calculated pair-wise for all transcripts. (D) Orthogonal validation of expression accuracy. Previous qPCR data for ~1000 mRNAs was compared with RNA-seq data [23].

Similar articles

See all similar articles

References

    1. Lucas MC, Jacobson JW and Giles NH. Characterization and in vitro translation of polyadenylated messenger ribonucleic acid from Neurospora crassa. J. Bacteriol. 1977;130, 1192–1198. - PMC - PubMed
    1. Sturani E, Costantini MG, Martegani E, and Alberghina L. Level and turnover of polyadenylate-containing ribonucleic acid in Neurospora crassa in different steady states of growth. Eur. J. Biochem. 1979; 99, 1–7. 10.1111/j.1432-1033.1979.tb13224.x - DOI - PubMed
    1. Aviv H and Leder P. Purification of biologically active globin messenger RNA by chromatography on oligothymidylic acid–cellulose. Proc. Natl. Acad. Sci. U.S.A. 1972;69, 1408–1412. 10.1073/pnas.69.6.1408 - DOI - PMC - PubMed
    1. Kuai L, Fang F, Butler JS, and Sherman F. Polyadenylation of rRNA in Saccharomyces cerevisiae. Proc. Natl. Acad. Sci. U.S.A. 2004;101, 8581–8586. 10.1073/pnas.0402888101 - DOI - PMC - PubMed
    1. Adiconis X, Borges-Rivera D, Satija R, DeLuca DS, Busby MA, Berlin AM, et al. Comparative analysis of RNA sequencing methods for degraded or low-input samples. Nat. Methods 2013;10, 623–629. 10.1038/nmeth.2483 - DOI - PMC - PubMed

Grant support

MAM acknowledges support from the Canadian Institutes of Health Research (FDN-143288) the BC Cancer Foundation, Genome Canada (212SEQ) and Genome British Columbia (202SEQ). We are grateful for support from the Canada Foundation for Innovation and the National Cancer Institute, National Institutes of Health, under Contract No. HHSN261200800001E. The content of this publication does not necessarily reflect the views of policies of the Department of Health and Human Services, nor does mention of trade names, commercial products, or organizations imply endorsement by the U.S. Government.
Feedback