We use cookies to improve your browsing experience and provide meaningful content. Read our cookie policy. Accept

Clontech Takara Cellartis

Home › Learning centers › Next-generation sequencing › Technical notes › RNA-seq › Full-length, single-cell, and ultra-low-input RNA-seq with UMIs

Technical notes

New products
Need help?
Contact Sales
SSmRNA + UMIs product information
Tech Note

UMI-based RNA counting and full-length transcriptome analysis from single-cell and ultra-low inputs

RNA sequencing (RNA-seq) captures the transcriptomic repertoire and quantifies the expression levels of genes. When performed on single cells (scRNA-seq), it is a powerful tool for characterizing cellular heterogeneity and understanding complex biological processes, including disease mechanisms.

Single-cell and ultra-low-input RNA-seq data are especially susceptible to errors and biases introduced during library preparation, resulting in higher variability. Unique molecular identifiers (UMIs) improve the data accuracy and reproducibility of ultra-low-input and scRNA-seq data. UMIs are short, random nucleotide sequences—barcodes—that uniquely tag individual molecules in a sample library before the amplification step. The UMIs enable elimination of PCR errors and duplicates, improve the accuracy of differential gene expression (DGE) measurements, and improve variant detection especially for rare mutations.

scRNA-seq end counting with UMIs is well-established for DGE measurements. However, the end-focused approach provides limited information about structural variations in the transcript. Full-length RNA-seq data reveals isoforms, SNPs, fusions, and other biologically relevant features. However, full-length RNA-seq workflows typically lack UMIs. It can also be challenging to incorporate UMIs during full-length RNA-seq library prep in a way that maintains yields, sensitivity, and reproducibility.

To address these limitations, we created SMART-Seq mRNA LP (with UMIs), abbreviated here as SSmRNA + UMIs. It is a strategic combination of our established single-cell/ultra-low-input RNA-seq kits: SMART-Seq mRNA and SMART-Seq mRNA LP (SSmRNA) with UMI-based RNA counting.

In this tech note, we demonstrate that SSmRNA + UMIs delivers the same high sensitivity and performance as the established SSmRNA kits. SSmRNA + UMIs excels at single-cell and ultra-low-input RNA-seq, outperforming the home-brew full-length RNA-seq method, Smart-seq2 (SS2). We show that UMIs improve RNA counting accuracy and increase sensitivity at single cell inputs. With our novel library prep method, automation-friendly workflow, and free data analysis tools, the kit provides a complete solution for full-length RNA-seq with end counting.

SMART-Seq technology + UMIs  

SSmRNA + UMIs, the latest improvement in SMART-Seq technology

SMART (Switching Mechanism At 5' end of RNA Template) technology was first described in Biotechniques in 2001 (Zhu et al. 2001) and powers the whole-transcriptome RNA-seq kits in our SMART-Seq portfolio. It leverages the template-switching activity of reverse transcriptases derived from Moloney Murine Leukemia Virus (MMLV) to efficiently capture the 5’ end of transcripts. It is capable of synthesizing full-length cDNA templates up to ~15 kb, incorporating anchor sequences to the ends of the cDNA templates for subsequent cDNA amplification.

Takara Bio introduced the first RNA-seq kit using SMART technology in 2011, and Figure 1 highlights kits in the SMART-Seq portfolio since 2015. These products improved sensitivity while optimizing for specialized applications, such as 3’ differential gene expression, high-throughput analysis, and single-cell RNA-seq specifically formulated for cells with ultra-low RNA content (e.g., peripheral blood mononuclear cells [PBMC], T cells, and B cells). The SMART-Seq mRNA kits have become the industry standard for ultra-low input RNA-seq.

Figure 1. The SMART-Seq portfolio includes specialized kits for low-input and single-cell RNA. The addition of UMIs to the original SMART-Seq mRNA LP kit provides higher accuracy in RNA counting. Now it is possible to detect rare transcripts and identify rare cell types with high confidence.

The new SSmRNA + UMIs kit produces final libraries that include 5′-UMI-tagged fragments and internal fragments lacking UMIs, spanning the entire transcript. These libraries enable a combination of full-length RNA-seq with quantitative RNA counting. The complete SMART-Seq chemistry with UMI incorporation is shown in Figure 2.

SSmRNA + UMIs chemistry

SSmRNA + UMIs chemistry

Figure 2. UMIs are now included in SMART-Seq technology for ultra-low amounts of RNA and single cells. Nontemplated nucleotides (depicted as XXXXX), added by SMARTScribe Reverse Transcriptase (RT) hybridize to the UMI-containing TSO. This provides a new template for the RT. The SMART adapters used for amplification during PCR—added by the oligo(dT) primer and TSO—are indicated in green. 5’ ends of the oligo-dT and SMART-Seq mRNA PCR Primers are chemically modified (depicted by black stars) to prevent the ligation of sequencing library adapters to DNA fragments carrying the SMART sequence during sequencing library construction. cDNA is enzymatically fragmented, stem-loop adapters (red and blue) ligated, and the molecules extended, all in a single tube. Extension produces linear molecules that are then amplified and indexed to generate Illumina-compatible libraries with unique dual indexes (UDIs; light gray). The final library includes UMI (yellow)-tagged fragments as well as internal fragments (that do not carry UMIs) spanning the whole gene body.

SSmRNA + UMIs combined with our free Cogent NGS data analysis tools offer a complete workflow from cDNA synthesis to analysis (Figure 3). Cogent NGS Analysis Pipeline (CogentAP) maps and reports RNA-seq data for full-length UMI-based or UMI-agnostic transcriptome analysis. Included are gene and transcript counting, gene fusion detection, and immune profiling analysis. The Cogent NGS Discovery Software (CogentDS), an interactive data visualization tool, uses output files from CogentAP for downstream cluster analysis.

Figure 3. SSmRNA + UMIs combined with Cogent NGS data analysis tools are a complete solution for full-length transcriptome analysis and RNA counting for single-cell and ultra-low-input RNA-seq. The kit's workflow goes from cDNA synthesis to library prep. Sequence on an Illumina instrument and use the resulting FASTQ files as input for our free analysis software. Output from CogentAP can be input for CogentDS to create Uniform Manifold Approximation and Projection (UMAP) and t-Distributed Stochastic Neighbor Embedding (t-SNE) plots.

Uncompromised performance with UMIs  

High-quality RNA-seq libraries with SSmRNA + UMIs for single-cell and ultra-low inputs

Once UMIs were included with the SSmRNA chemistry, it was important to test whether this addition affected the sensitivity of the RNA-seq process.

We compared RNA-seq data from libraries generated using SSmRNA + UMIs with SSmRNA for total RNA (Figure 4) and single-cell input (Figure 5). The data showed comparable gene counts and read distribution metrics, indicating data quality was not compromised by including UMIs.

Figure 4. SSmRNA + UMIs is as sensitive as SSmRNA. SMART-Seq mRNA LP (with UMIs) and SMART-Seq mRNA LP were each used to create sequencing libraries from K562 RNA (10 pg) or universal human reference RNA (100 pg, 1 ng, 10 ng, and 100 ng). For all conditions, n = 3. Libraries were sequenced on the Illumina MiSeq. Sequencing data from K562 samples were downsampled to 1.8 million reads, while those from all UHR samples were downsampled to 1.0 million reads. Data were processed using CogentAP.

Figure 5. SSmRNA + UMIs has uncompromised sensitivity with single-cell inputs. Single peripheral blood mononuclear cells (PBMCs) were sorted into plates. Sequencing libraries were produced using either SSmRNA or SSmRNA + UMIs. Libraries were pooled, sequenced, and downsampled to 900,000 reads for analysis. Data was processed using CogentAP.

SSmRNA + UMIs delivers superior sensitivity and reproducibility to SS2

Comparing the performance of SSmRNA + UMIs to SS2 showed that SSmRNA + UMIs provided higher gene counts (11,082) than SS2 (9,964), demonstrating greater sensitivity (Figure 6). SSmRNA + UMIs also showed lower technical variability (avg. R2 = 0.963) compared to SS2 (avg. R2 = 0.92) demonstrating superior reproducibility of the validated SSmRNA + UMIs kit compared to the SS2 protocol (Figure 7).

Unlike SSmRNA + UMIs, a kit with reagents that have been optimized and validated to work together, SS2 is a homebrew protocol. This entails using reagents from various vendors—the compatibility and performance of the reagents could vary—necessitating user optimization for optimal performance.

Figure 6. SSmRNA + UMIs outperforms SS2 in terms of sensitivity. cDNA and sequencing libraries (n = 3) were generated from 10 pg UHR RNA using SSmRNA + UMIs or the SS2 method. SS2 libraries were prepared using the Nextera XT DNA Library Prep Kit (Illumina). After pooling and sequencing, all samples were downsampled to 1.2 million reads. Data was processed using CogentAP.

Figure 7. SSmRNA + UMIs outperforms SS2 in terms of reproducibility. Our free bioinformatic software, Cogent AP, was used to plot absolute read counts and perform regression analysis for each SS2 and SSmRNA + UMIs replicate pair. The R2 values were consistently higher for SSmRNA + UMIs. An example XY plot for each method is shown.

Accurate RNA counting with UMIs  

UMIs increase accuracy of RNA quantitation by deduplication through UMI-collapse

Single-cell RNA-seq applications start with very low RNA inputs, with sensitivity and accuracy becoming more of a challenge. Therefore, incorporating UMIs to control errors and amplification artifacts can dramatically improve overall data accuracy. Deduplication via UMI collapse removes biological duplicate reads from the sequencing data, thereby increasing the accuracy of transcript quantitation.

RNA counting accuracy with UMIs was tested using External RNA Controls Consortium (ERCC) RNA Spike-In controls (Figure 8). Increasing amounts of ERCC molecules were added to ultra-low amounts of bulk RNA (10 pg UHR) or single-cell (PBMC) inputs, and RNA-seq libraries were prepared using SSmRNA + UMIs. Read counts for measured versus absolute number of input ERCC molecules showed higher concordance (higher R2) after UMI collapse for both bulk and single-cell inputs.

The increase in R2 values after UMI collapse demonstrates that incorporating UMIs in the SSmRNA workflow significantly improves transcript measurement accuracy by eliminating PCR duplicates.

Figure 8. Using UMIs increases the accuracy of RNA quantitation for bulk RNA or single-cell inputs. 10 pg of input RNA or RNA from a single cell was spiked with ERCC control RNA. Libraries were prepared using the SSmRNA + UMIs method. After sequencing, CogentAP was used to plot TPM or CPM against the absolute number of input molecules from the ERCC spike-in. Subsequent regression analysis showed a significant increase in the R2 with the collapse of UMIs.

Automation and miniaturization  

SSmRNA + UMI is compatible with automation and miniaturization

Research groups with high-throughput needs may require automated preparation of sequencing libraries. Automation reduces hands-on time, improves processing reproducibility, and minimizes handling errors. Miniaturization dramatically reduces overall sample processing costs. Liquid handling systems such as the Formulatrix MANTIS Liquid Dispenser and SPT Labtech mosquito HV offer the ability to automate the library preparation workflow and perform reactions at smaller volumes.

We tested these commonly used platforms in the preparation of sequencing libraries from 10 pg of mouse brain RNA. The SSmRNA + UMIs reactions proceeded at fractions of the full volume used in manual preparation. Figure 9 shows the resulting RNA subtype representation and gene counts. There is a high concordance in read distributions and sensitivity across all volumes.

Therefore, we can confirm that SSmRNA + UMIs provides data of consistently high quality even when adapted to automated liquid handling systems.

Figure 9. SSmRNA + UMIs is compatible with automation. Mouse brain control RNA (10 pg, n = 8) was used as input for the SSmRNA + UMIs workflow at full-, half-, and quarter-volume reactions on the MANTIS Liquid Dispenser. Libraries were pooled, sequenced, and downsampled to 1.3 million reads. Identical RNA inputs were also processed at one-eighth volume on the mosquito HV liquid handler, or manually on the benchtop. These libraries were then pooled, sequenced, and downsampled to 620,000 reads. All analysis was performed using CogentAP.

Conclusions and references  

Conclusions

With the introduction of SMART-Seq mRNA LP (with UMIs), we now offer a complete, robust RNA counting strategy for full-length transcriptome studies across a range of inputs, from single-cell to 100 ng RNA.

The UMIs allow for RNA counting without compromising data quality and provide greater confidence for measuring transcript abundance. They enhance the sensitivity and accuracy of DGE analysis, especially for ultra-low- and single-cell-input libraries, which suffer from PCR duplicates and biases due to higher amplification cycles during library preparation.

SSmRNA + UMIs demonstrated higher sensitivity and reproducibility as compared to the SS2 homebrew protocol, and comparable performance to SSmRNA for single-cell and ultra-low RNA inputs.

This workflow is compatible with automation systems, with high sensitivity being maintained in miniaturization. The kit forms part of a complete solution, from sample to data, using our free-to-use Cogent NGS data analysis tools.

References

Zhu, Y. Y., Machleder, E. M., Chenchik, A., Li, R., and Siebert, P. D. Reverse Transcriptase Template Switching: A SMART™ Approach for Full-Length cDNA Library Construction. BioTechniques 30, 892–897 (2001). https://doi.org/10.2144/01304pf02

Takara Bio USA, Inc.
United States/Canada: +1.800.662.2566 • Asia Pacific: +1.650.919.7300 • Europe: +33.(0)1.3904.6880 • Japan: +81.(0)77.565.6999
FOR RESEARCH USE ONLY. NOT FOR USE IN DIAGNOSTIC PROCEDURES. © 2023 Takara Bio Inc. All Rights Reserved. All trademarks are the property of Takara Bio Inc. or its affiliate(s) in the U.S. and/or other countries or their respective owners. Certain trademarks may not be registered in all jurisdictions. Additional product, intellectual property, and restricted use information is available at takarabio.com.

©2024 Takara Bio Inc. All Rights Reserved.

Region - North America Privacy Policy Terms and Conditions Terms of Use

Top