Increasing the Sensitivity of Transcriptome Profiling in Eukaryote and Blood samples by depleting abundant RNAs Lynne M. Apone, Deyra N. Rodríguez, Kaylinnette Pinet, Bradley W. Langhorst, Jian Sun, Brittany Sexton, Anetta Nowosielska-Vecchietti, Timur Shtatland, Christine J. Sumner, Fiona J. Stewart, Eileen T. Dimalanta and Theodore B. Davis. | New England Biolabs, Inc.

Introduction Results

The large dynamic range of transcript expression within a total RNA sample presents a challenge in whole-transcriptome sequencing. Efficient rRNA depletion across species and a wide input range Efficient rRNA depletion in degraded

Highly expressed transcripts with minimal biological interest can dominate readouts, masking detection of more informative lower abundance A. Mouse Rat FFPE RNA 100% 100% 100% 11 99 99 99 10 99 99 99 9 99 97 99 Human Liver FFPE Total RNA 90% 90% 90% B. 90 91 transcripts. Here, we present a method to enrich for RNAs of interest by eliminating unwanted RNAs before sequencing. This method is 89 100 80% 80% 80% based on hybridization of probes to the targeted RNA and subsequent enzymatic degradation of the selected RNAs. 70% 70% 70% 90 95 60% 60% 60% 80 72 We optimized this method to remove cytoplasmic and mitochondrial ribosomal RNA (rRNA) from human, mouse, and rat total RNA 50% 50% 50% 70 * Reads (%) Reads Reads (%) Reads 40% 40% 40% 60 Reads (%) Reads 30% 30% 30% 50 samples. This includes removal of the internal and external transcribed spacers (ITS & ETS) in human primary rRNA transcripts, which 20% 20% 20% 40 33 10% 10% 10% 0.6 0.6 0.9 (%) rRNA Reads 1.2 1.1 1.3 0.9 2.9 1.0 30 accounts for up to ~10% of transcripts in certain sample types. We have also incorporated new enzymes and streamlined the method to 0% 0% 0% 1ug 100ng 10ng 1ug 100ng 10ng 1ug 100ng 10ng 20 5 make it more robust and specific. Additionally, and to address the presence of highly abundant transcripts in blood samples, we Undepleted Depleted Undepleted Depleted Undepleted Depleted 10 4 rRNA RNA of Interest rRNA RNA of Interest rRNA RNA of Interest 0 expanded the probe set to remove adult, fetal and embryonic globin transcripts from blood RNA from which globin can constitute up to ~60% 100ng 10ng Undepleted Depleted of mRNA transcripts. Transcript expression correlation is maintained after depletion NEBNext Illumina Using strand-specific RNA sequencing we measured depletion efficiency, library complexity and transcript expression before and after Figure 1. Consistent and efficient depletion across species and C. Human Mouse Rat inputs while maintaining transcript expression of non-targeted depletion. RNAs. The NEBNext rRNA Depletion Kit v2: A) efficiently removes rRNA from human, mouse and rat total RNA enabling a higher percentage (>97%) of reads to map to RNAs of interest; B) its Methods superior at depleting rRNA in FFPE (degraded) samples and offers the flexibility of lower total RNA input amounts. (*) Input not (ReadCounts, log 10) NEBNext Depletion Workflow Targeted RNA recommended by Illumina; C) does not disturb the expression of non-targeted RNA species resulting in high transcript abundance R 2=0.96 R 2=0.92 R 2=0.99 correlation between undepleted (no treatment) and depleted

NEBNext rRNA Depletion Kit v2 (h/m/r) Undepleted transcripts. GENCODE v27 transcript abundances were estimated Depleted (Read Counts, log 10) 2 Cytoplasmic rRNA: 28S, 18S, 5.8S, 5S, 5’ and 3’ human ETS, ITS1, ITS2 using Salmon [2]. Read Counts and R values for the linear fit are Mitochondrial rRNA:12S, 16S shown.

NEBNext Globin & rRNA Depletion Kit (h/m/r) Depletion of globin and ribosomal RNA enriches for RNAs of interest across species Cytoplasmic and Mitochondrial rRNA (as above) Globin mRNA: Adult: alpha (HBA 1,2), beta (HBB), delta (HBD), mu (HBM) Fetal: gamma (HBG1/2) Figure 2. Highly efficient depletion of globin and rRNA Embryonic: epsilon (HBE1), theta (HBQ1), zeta (HBZ) The NEBNext Globin and rRNA Depletion Kit efficiently removes rRNA and globin RNA from human, mouse and rat total RNA enabling a higher percentage (>95%) of reads to map to RNAs of interest. Here shown depletion on 1ug of Total RNA. High Experimental Design and Analysis depletion efficiency is also observed with 100ng and 10ng of input RNA (not shown). • Human, Mouse and Rat Universal Reference total RNA (1ug, 100ng, 10ng) and Human Liver FFPE total RNA (100ng, 10ng) were depleted of rRNA using the NEBNext Depletion Kit v2 (E7400) or the TruSeq Stranded Total RNA Gold depletion reagents (FFPE sample). Conclusions • Human, Mouse and Rat whole blood total RNA (1ug, 100ng, 10ng) was depleted of • The NEBNext Depletion Workflow is a quick and effective method to deplete unwanted RNA species for RNA-Seq. rRNA and globin mRNA transcripts using the NEBNext Globin & rRNA Depletion Kit • The workflow is suitable for intact and degraded RNA. (E7750). • The reagents are compatible with any downstream NGS library prep workflow, and amenable to automation. • RNA-seq libraries were prepared using the NEBNext Ultra™ II Directional RNA Library Prep Kit for Illumina® (NEB #E7760) followed by paired-end sequencing on • The NEBNext Depletion Kit v2 and the NEBNext Globin & rRNA Depletion Kit are highly effective in human, mouse and rat. a NextSeq ® instrument. • Compatible with a broad input range: 10ng - 1ug Total RNA. • Abundance of non-targeted transcripts is maintained after depletion and across range of inputs. Total RNA is hybridized with probes targeting unwanted, abundant RNAs (e.g. • Reads were identified as ribosomal or globin using Mirabait (6 or more, 25-mers) ribosomal RNA and globin mRNA), followed by RNase H digestion to degrade [1]. rRNA sequences (28S, 18S, 5.8S, 5S 12S, 16S, ITS, ETS) and globin mRNA • The removal of abundant transcripts using this simple and efficient method greatly increases the sensitivity of RNA-seq studies by targeted RNA. Finally, DNA probes are digested with DNase I, and the reaction sequences (HBA1/2, HBB, HBG1/2, HBD, HBE1, HBZ, HBM, HBQ1) were used as enabling the detection of lower abundance transcripts and true biological variations. is cleaned using magnetic beads. The entire workflow can be done in <2 hours baits. rRNA or globin remaining after depletion were calculated by dividing matched with only 8 minutes of hands-on time. Ribosomal RNA depletion can be • This method is amenable to high throughput sample preparation and robotic automation for easy implementation in a clinical setting. immediately followed by RNA-seq library preparation. reads by total number of reads passing instrument quality filtering. References 1. http://doi.org/10.1101/gr.1917404 2. http://doi.org/10.1101/021592