high throughput single cell rna-seq · n. shakir, m. breteau and h. johnson blacktrace holdings...
TRANSCRIPT
High Throughput Single Cell RNA-SEQN. Shakir, M. Breteau and H. JohnsonBlacktrace Holdings Ltd, United Kingdom
IntroductionCell encapsulation within microfluidic droplets is an exciting, cutting edge technique that, for the first time, enables the analysis of thousands or even millions of single cells. This unprecedented analytical power enables the discovery of previously unidentified or cryptic cell types in a tissue of interest and also the characterisation of individual cells within biologically important processes.
The Drop-Seq by Macosko et al. method, encapsulates a single cell alongside a single bead within a droplet. These beads have surfaces coated with DNA-oligos allowing for the capture of polyadenylated mRNA which is subsequently reverse transcribed into cDNA. Each bead also carries a unique DNA barcode that identifies not only individual cells but also individual mRNA transcripts derived from any given cell. The Dolomite Bio Nadia Instrument is designed to perform Drop-seq in high throughput to analysis of up to 8 single cell samples, in parallel in under 20 mins, using a fully automated process. Over 50,000 single cells can be captured per cartridge in a run. The fully automated Nadia instrument guides users through all relevant steps of the experiment via an easy-to-use touchscreen interface.
Encapsulation ResultsEven distribution of cells and beads in the Nadia instrument is maintained by the gentle stirring. To demonstrate that the stirring employed exerted no adverse effects, cellular viability was tested using 0.4% Trypan Blue Solution. The tested sample contained MEF cells that had been stirred in the Nadia Instrument for 30 min at 4 ˚C. Compared to cells that had been stored on ice for 30 min we did not observe differences in viability after stirring. For the Drop-seq run on the Nadia instrument a 1:1 mixture of both Human HEK and mouse 3T3 cells were encapsulated with barcoded mRNA capture beads. Cells were encapsulated at a ratio of one cell per 20 droplets. The number of beads per droplets was determined after the run by a count of the numbers of beads encapsulated within a single droplet and this was performed across 450 droplets. Beads had been suspended at a theoretical encapsulation rate of 1 bead per 10 droplets (or 10 % of droplets). 50 droplets out of 450, or 11.0% of droplets contained a single bead. This was in agreement with theoretical expectations. The average droplets size was 85 μm with a CV of 7 %.
Downstream AnalysisSTAMPs were identified from a pool of 2000 amplified beads obtained after encapsulation of both human and mouse cells using the Nadia. Only a fraction of the beads had been encapsulated within droplets that contained only a single bead and cell, whilst the majority may have had contact with RNA from the extracellular milieu. In the Figure, the cell barcodes were arranged in descending order of reads and visualised alongside the cumulative fraction of reads in a knee-plot. An inflection-point can be observed at 114 and this corresponds well to the 100 cells predicted by Poisson statistics for 2000 beads. When compared with published data from Macosko and co-workers, the numbers of genes expressed by cells studied using Nadia were highly consistent at a given depth of NGS-analysis. The Barnyard plot was used to visualize the species specificity of individual STAMPs. For this dataset 72 STAMPs were associated with only mouse transcripts, 41 with only human transcripts and one single STAMP carried transcripts from both species which equated to a doublet rate of 1.8 %.
ConclusionThrough this work we demonstrate that the Nadia Instrument can successfully encapsulate cells alongside barcoded beads within droplets containing lysis buffer. The device produced monodispersed droplets with an approximate cell capture rate of 10%. It was also shown that gentle stirring of beads and cells on the Nadia Chips maintained beads and cells in suspension, facilitated encapsulation and did not damage cells. A high quality single cell DNA library was obtained from use of the Nadia Instrument and this gave rise to a dataset which after NGS-analysis was deemed to be of a high quality. In conclusion, the RNA-Seq application using the Drop-seq protocol has been demonstrated with good results using Nadia Instrument.
[1] Macosko E., et al. “Highly Parallel Genome-wide Expression Profiling of Individual Cells Using Nanoliter Droplets.” Cell 161:1202.
AAA
CCCGGG
123456
TTT
TTTAAA
Beads
5’ 3’
5’TSO
Cells
Oil
cDNA Library Primer Binding Site
Cell Barcode Unique Molecular Identifier
Structure of the Barcoded Oligonucleotide
TTTT T T
TTT
Oil
BC
5’
3’
Droplet
Bead
Oil Oil
Cell
BarcodeBead Barcode
mRNA
Droplet
AAA
CCCGGG
123456
TTT
TTTAAA
Beads
5’ 3’
5’TSO
Cells
Oil
cDNA Library Primer Binding Site
Cell Barcode Unique Molecular Identifier
Structure of the Barcoded Oligonucleotide
TTTT T T
TTT
Oil
BC
5’
3’
Droplet
Bead
Oil Oil
Cell
BarcodeBead Barcode
mRNA
Droplet
AAA
CCCGGG
123456
TTT
TTTAAA
Beads
5’ 3’
5’TSO
Cells
Oil
cDNA Library Primer Binding Site
Cell Barcode Unique Molecular Identifier
Structure of the Barcoded Oligonucleotide
TTTT T T
TTT
Oil
BC
5’
3’
Droplet
Bead
Oil Oil
Cell
BarcodeBead Barcode
mRNA
Droplet
AAA
CCCGGG
123456
TTT
TTTAAA
Beads
5’ 3’
5’TSO
Cells
Oil
cDNA Library Primer Binding Site
Cell Barcode Unique Molecular Identifier
Structure of the Barcoded Oligonucleotide
TTTT T T
TTT
Oil
BC
5’
3’
Droplet
Bead
Oil Oil
Cell
BarcodeBead Barcode
mRNA
Droplet
AAA
CCCGGG
123456
TTT
TTTAAA
Beads
5’ 3’
5’TSO
Cells
Oil
cDNA Library Primer Binding Site
Cell Barcode Unique Molecular Identifier
Structure of the Barcoded Oligonucleotide
TTTT T T
TTT
Oil
BC
5’
3’
Droplet
Bead
Oil Oil
Cell
BarcodeBead Barcode
mRNA
Droplet
AAA
CCCGGG
123456
TTT
TTTAAA
Beads
5’ 3’
5’TSO
Cells
Oil
cDNA Library Primer Binding Site
Cell Barcode Unique Molecular Identifier
Structure of the Barcoded Oligonucleotide
TTTT T T
TTT
Oil
BC
5’
3’
Droplet
Bead
Oil Oil
Cell
BarcodeBead Barcode
mRNA
Droplet
AAA
CCCGGG
123456
TTT
TTTAAA
Beads
5’ 3’
5’TSO
Cells
Oil
cDNA Library Primer Binding Site
Cell Barcode Unique Molecular Identifier
Structure of the Barcoded Oligonucleotide
TTTT T T
TTT
Oil
BC
5’
3’
Droplet
Bead
Oil Oil
Cell
BarcodeBead Barcode
mRNA
Droplet
Cell: 1 2 ... N
GENE 1 1 2 14
GENE 2 4 27 8
GENE 3 0 0 1
. . . . .
. . . . .
. . . . .
GENE M 6 2 0
AAA
CCCGGG
123456
TTT
TTTAAA
Beads
5’ 3’
5’TSO
Cells
Oil
cDNA Library Primer Binding Site
Cell Barcode Unique Molecular Identifier
Structure of the Barcoded Oligonucleotide
TTTT T T
TTT
Oil
BC
5’
3’
Droplet
Bead
Oil Oil
Cell
BarcodeBead Barcode
mRNA
Droplet
NGS sequencing
Library amplification
Downstream First strand cDNA synthesis
Cell lysis and mRNA capture
Nadia Instrument Microfluidic compartmentalization of cells and barcoded mRNA capture beads
Upstream Sample preparation
Bioinformatics pipeline
Data visualization
1
2
3
Hum
an tr
ansc
ripts
Mouse transcripts
Mouse Human Mixed
72411
0 500 10000 15000 20000 25000 30000
020
000
4000
060
000
8000
010
000
1200
014
000
Cum
ulat
ive
frac
tion
of re
ads 1.00
0.75
0.50
0.25
0.00
STAMPS each associated with a unique cell barcode were sorted in descending order based on corresponding number of NGS-reads
0 200 400 600 800
Med
ian
num
ber o
f Gen
es a
ssoc
iate
d w
ith a
STA
MP
020
000
4000
060
000
8000
0
Depth of NGS-analysisMean number of NGS-Reads associated with a
STAMP (x 1000)
HEK293T (Human) 3T3 (Mouse)
Data from Macosko et al.,
100 STAMP-dataset
Nadia_2000 Nadia_4000 Data generated using Nadia.
0 500 1000 1500
Figure 3. (A) This knee-plot shows the cell barcodes in descending order of reads and visualised alongside the cumulative fraction of reads. (B) Numbers of genes expressed by cells studied using Nadia compared with published data from Macosko and co-workers.
Figure 2. Species mixing experiment results, indicating amount of transcripts captured and low doublet rates of mouse and human samples.
Figure 1. Droplets produced through an scRNA-seq run on the Nadia Instrument