CEL-Seq2: sensitive highly-multiplexed single-cell RNA-Seq.

Hashimshony T, Senderovich N, Avital G, Klochendler A, de Leeuw Y, Anavy L, Gennert D, Li S, Livak KJ, Rozenblatt-Rosen O, Dor Y, Regev A, Yanai I. · 2016

Read CEL-Seq2: sensitive highly-multiplexed single-cell RNA-Seq. by Hashimshony T, Senderovich N, Avital G, Klochendler A, de Leeuw Y, Anavy L, Gennert D, Li S, Livak KJ, Rozenblatt-Rosen O, Dor Y, Regev A, Yanai I. (2016) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/europmc-pmc4848782.txt.

Introduction

Tamar Hashimshony and co-authors presented CEL-Seq2 in Genome Biology in 2016, an upgraded method to read gene activity in single cells. Single-cell work needs a method that is sensitive, exact, and repeatable, and cheap enough to run on many cells at once. CEL-Seq2 triples the sensitivity of the group's earlier method while cutting cost and hands-on time.

The team also ran the method on a chip system that traps single cells, its first on-chip barcoding use, and caught gene shifts as mouse cells moved through the division cycle. Against a rival method, CEL-Seq2 spotted more faint genes. The gains suit large surveys that sort cells by their active genes.

Catalog record

CREATORHashimshony T, Senderovich N, Avital G, Klochendler A, de Leeuw Y, Anavy L, Gennert D, Li S, Livak KJ, Rozenblatt-Rosen O, Dor Y, Regev A, Yanai I.
YEAR2016
COLLECTIONPAPERS
FORMATTXT
FILE SIZE24 KB
LICENSECC BY 4.0
RIGHTSCC BY
SOURCEEurope PMC

Source: https://europepmc.org/articles/PMC4848782

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

CEL-Seq2: sensitive highly-multiplexed single-cell RNA-Seq By Tamar Hashimshony, Naftalie Senderovich, Gal Avital, Agnes Klochendler, Yaron de Leeuw, Leon Anavy, Dave Gennert, Shuqiang Li, Kenneth J. Livak, Orit Rozenblatt-Rosen, Yuval Dor, Aviv Regev Published in: Genome Biology DOI: 10.1186/s13059-016-0938-8 ABSTRACT Single-cell transcriptomics requires a method that is sensitive, accurate, and reproducible. Here, we present CEL-Seq2, a modified version of our CEL-Seq method, with threefold higher sensitivity, lower costs, and less hands-on time. We implemented CEL-Seq2 on Fluidigm’s C1 system, providing its first single-cell, on-chip barcoding method, and we detected gene expression changes accompanying the progression through the cell cycle in mouse fibroblast cells. We also compare with Smart-Seq to demonstrate CEL-Seq2’s increased sensitivity relative to other available methods. Collectively, the improvements make CEL-Seq2 uniquely suited to single-cell RNA-Seq analysis in terms of economics, resolution, and ease of use. Electronic supplementary material The online version of this article (doi:10.1186/s13059-016-0938-8) contains supplementary material, which is available to authorized users. BACKGROUND Single-cell transcriptomics is a transformative method with tremendous potential to illuminate the complexities of gene regulation. Single-cell RNA-Seq was first introduced by Tang et al. [1], using a polyT primer with an anchor sequence to select for the cell’s mRNA. After polyadenylation of the resulting cDNA, a second polyT primer with a different anchor is used to obtain double stranded DNA, which is then PCR-amplified. Each sample is individually converted to a library for sequencing. The STRT method introduced early barcoding at the reverse transcription stage [2], thereby enabling highly-multiplexed analyses, and adapted a template switching mechanism based on the ability of the reverse transcriptase to tag the end of the cDNA [3], eliminating the need for the polyadenylation reaction. Smart-Seq [4, 5] used the same template switching mechanism as STRT, but without the early barcoding. Each sample is processed individually, and the reaction was optimized for full transcript sequencing. The CEL-Seq [6] method is the first method to use in vitro transcription (IVT) for the amplification, thereby eliminating the requirement for a template-switch step which is thought to reduce efficiency. We use early barcoding, enabling highly-multiplexed analysis, and 3′ end tagging enabling accurate estimation of expression levels without having to account for gene length and with fewer sequencing reads required. Here we introduce CEL-Seq2, which has been optimized for higher sensitivity, less hands-on time, and lower price. We show that CEL-Seq2 works well on different platforms, and compare it to previously published methods. RESULTS AND DISCUSSION CEL-Seq2 is optimized for higher sensitivity Recent adaptations of CEL-Seq [7, 8] integrated unique molecular identifiers [9, 10] (UMI) into the CEL-Seq primer, enabling each reverse-transcribed mRNA to be counted precisely once. Estimating CEL-Seq’s sensitivity as the fraction of ERCC spike-ins [11] transcripts detected using such UMIs, we and others [7] computed CEL-Seq’s efficiency at ~6 %. This may be an underestimate, however, because comparison with smFISH indicates threefold higher sensitivity [7]. Seeking to improve CEL-Seq’s efficiency, we introduced several changes, summarized in Fig. 1a.Fig. 1Changes introduced to the protocol. a An outline of the CEL-Seq2 method is shown with the steps modified from the original CEL-Seq indicated in red. b Distribution of the reads in two libraries prepared with or without ligation from the same amplified RNA of ten replicates of 100 pg clean RNA. c, d Number of transcripts (c) and genes (d) detected. Error bars indicate standard error First, we sought to increase the efficiency of the reverse transcription (RT) reaction by shortening the CEL-Seq primer from 92 to 82 nucleotides, despite the addition of six UMI nucleotides. This was done by reducing the length of the barcode from eight to six nucleotides, as well as shortening the T7 promoter and the Illumina 5’ adaptor. Use of the shortened primers indeed improved the sensitivity to 10.6 %, detecting more transcripts (Additional file 1: Figure S1). In this analysis on 100 pg of RNA, the number of detected genes also increased, though not significantly (Additional file 1: Figure S1a), which likely reflects that most of the additionally identified transcripts are of genes already detected using the longer primer. We next optimized the conversion of RNA to dsDNA by testing alternative commercially available reverse transcriptases for cDNA synthesis and polymerases for second-strand synthesis. We found that SuperScript II for the RT step (Additional file 1: Figure S1d) provided a major improvement. For second strand synthesis, the differences were less pronounced, but the polymerase and other components from the SuperScript II Double-Stranded cDNA Synthesis Kit were better than its competitors (Additional file 1: Figure S1e). We also modified our method of dsDNA and aRNA clean-up from column to beads, which provided a threefold gain in yield (Additional file 1: Figure S1d). While CEL-Seq was originally implemented using the Ambion MessageAmp II aRNA Amplification Kit, these changes led to a kit-free (and therefor cheaper) process because the reagents of the SuperScript II Double-Stranded cDNA Synthesis Kit may be purchased separately. Ligation-free library preparation improves read mapping In the original CEL-Seq protocol, the aRNA is converted to a library compatible with Illumina sequencing by ligating the second adaptor [6]. Following conversion to cDNA with RT, a few PCR cycles completed the attachment of Illumina adaptors. The ligation step is not efficient, however, and introduces primer dimers that interfere with sequencing. CEL-Seq2 remedies this by inserting the Illumina adaptor directly at the RT step as a 5′-tail attached to a random hexamer (Fig. 1a, change 5), thus eliminating the ligation step. We prepared a library introducing the Illumina adaptor by ligation or using the random hexamer (no ligation) from the same amplified RNA. Sequencing of the “no ligation” libraries yields 93.8 % mapping of the reads with barcodes, an improvement from 60.9 % in the ligated library (Fig. 1b). This modification also led to the identification of more genes and transcripts (Fig. 1c, d), suggesting that the removal of the ligation step significantly increased the sensitivity. To control for sequencing depth, we sub-sampled 300,000 reads from each sample, and obtained similar results (Additional file 2: Figure S2a, b). In addition, this modification reduces the hands-on time and cost of library preparation and it alleviates the need for Illumina’s TruSeq Small-RNA kit. CEL-Seq2 is compatible with different platforms Our implementation of CEL-Seq for individual tubes is easily scaled-up to plates and could be performed with robotic liquid handlers. To further improve the efficiency, we sought to implement CEL-Seq2 on the Fluidigm C1, a nanoliter automatic microfluidic instrument. After capture of individual cells, the C1 loads the individually barcoded CEL-Seq2 primers from the outlet wells, lyses the cells chemically (rather than our traditional lysis by freezing), and performs an RT reaction followed by second-strand synthesis. In place of a cDNA clean-up, the second strand enzymes are heat inactivated. IVT then occurs for each sample individually. The amplified RNA is harvested from the C1 chip and pooled to a single sample from which a library is prepared. We performed CEL-Seq2 both manually and using the C1 on mouse fibroblast cells carrying a CyclinB1-GFP fusion reporter [12] and compared this to data obtained using the original protocol (with the single modification that the primers contained UMIs). We processed 24 cells using the original CEL-Seq protocol and 20 cells using CEL-Seq2, as well as

Read the complete text