A prerequisite for CIM-seq is the dissociation of cell lines or tissues into cell suspensions containing both single cells and multiplets. Dissociation procedure will have to be adjusted to the tissue and species of interest and a general method will therefore not be included here. For dissociation of murine lung, spleen, small intestine and colon as performed in the CIM-seq manuscript, please see the methods section of the CIM-seq manuscript.
We have divided this procedure section into three subsections:
A) Cell sorting and library preparation
B) Feature selection and clustering (Seurat)
C) Multiplet deconvolution
The library preparation section will describe Smartseq2 (well based) library preparation. For 10x Genomics, consult the manufacturers instructions for library prep found at: https://support.10xgenomics.com/single-cell-gene-expression/library-prep/doc/user-guide-chromium-single-cell-3-reagent-kits-user-guide-v3-chemistry.
A) Cell sorting and library preparation using Smartseq2:
1) Prepare lysis buffer mix (2µl per well):
NOTE: Reagents are prepared on ice, working quickly. ERCC is stored in single-use aliquots at -80 °C , thawed on ice and added last.
Reagent - μl per reaction
H20 - 1.31μl
RNase Inhibitor - 0.05μl
ERCC (1:600 000) - 0.05μl
Triton-X100 (10% solution) - 0.04μl
10mM dNTP - 0.5μl
100uM dT primer - 0.05μl
Total - 2μl
Add 2µl lysis buffer mix to each well. Cover with appropriate lids. Spin down.
Snap freeze on dry ice. Store until use at -80 °C
2) Sort single cells and multiplets (aggregates of multiple cells)
Sort into wells containing 2 µl lysis buffer mix.
Multiplets can be discerned from singlets by gating on the basis of FSC-W (Forward scatter - Width) and FSC-H (Forward scatter - Height) (see Figure 1).
Following sort, immediately seal with appropriate seals (approved for -80°C) and centrifuge at 2000 x g, 4°C, 5 min. Snap freeze on dry ice. Store until use at -80 °C.
3) Primer annealing
NOTE: Perform steps 3 and 4 as quickly as possible in order to prevent RNA degradation
Thaw plate and spin down briefly. Incubate in thermocycler at 72°C for 3 min to anneal primers. Place on ice immediately.
4) Reverse transcription
Prepare RT master-mix:
Reagent - Reagent volume
SmartScribe - 0.475μl
RNase Inhibitor - 0.125μl
5x First Strand buffer - 1μl
DTT (100mM) - 0.25μl
Betaine (5M) - 1μl
MgCl2 (1M) - 0.03μl
TSO (100uM) - 0.05μl
H20 - 0.07μl
Total - 3μl
Dispense 3µl per well, cover plate with new film and spin down. Incubate in thermocycler:
42°C for 90min.
70°C for 5 min.
4°C hold.
4) cDNA preamplification:
Prepare cDNA preamplification mix:
Reagents - Reagent volume
H2O - 1.0688μl
Kapa HiFi HotStart ReadyMix (2x) - 6.25μl
IS_PCR primer (10uM) - 0.125μl
Lambda Exonuclease - 0.05625μl
Total - 7.5μl
Dispense 7.5 µl per well. Total reaction volume will be 12.5µl. Spin down and cover with new lid. Incubate in thermocycler with the following program:
1 cycle:
37°C for 30 min.
95°C for 3 min.
Repeat 18-24 cycles (depending on initial RNA concentration):
98°C for 20 sec.
67°C for 15 sec.
72°C for 4 min.
1 final elongation cycle:
72°C 5 min.
4°C min. hold
5) cDNA cleanup
For cDNA cleanup we prepare SPRI-beads in 20% PEG-8000 solution as described in: https://openwetware.org/wiki/SPRI_bead_mix#Ingredients_for_50_mL_2
Using 20% SPRI-bead solution:
i. Add 0.7x the reaction volume of SPRI beads per well. Mix well by pipetting. (i.e 8.75 µl SPRI-bead solution for 12.5 µl reaction volume)
ii. Incubate 5 min. Room temperature
iii. Place on magnetic stand for 3 min.
iv. Carefully remove supernatant
v. Add 40 µl 80% EtOH and incubate 30 sec.
vi. Remove EtOH (without disturbing the beads)
vii. Wash again with EtOH. Make sure to remove as much ethanol as possible.
viii. Allow beads to air-dry for 10 - 15 min.
ix. Remove plate from magnetic stand
x. Elute beads in 15 µl EB or TE buffer. Mix well by pipetting
xii. Incubate 5 min. Room temperature
xiii. Place on magnetic plate for 3 min.
xiv. Optional: Carefully transfer supernatant to a new plate
6) cDNA quantification
In order to normalize cDNA amounts, measure concentration of randomly selected wells using Qubit HS dsDNA, adapted to a 96-well plate reader (May also be performed using standard Qubit protocol). Concentrations should be sampled both from wells containing single cells and multiplets in order to normalize separately.
i. Add 97 µl of 1X Qubit HS dsDNA solution to a flat-bottom, black plate
ii. Add 3 µl of cDNA sample
iii. Add Standards (NOTE: We make a 8-step ladder from 0ng/µl --> 10ng/µl Qubit Standard DNA in TE buffer)
iv. Read in plate reader using 485nM excitation/528nm emission
v. Calculate average well cDNA concentration for single cells and multiplets respectively
7) (optional) cDNA quality control
Using Agilent HS 5000 DNA chips (or equivalent) to profile cDNA fragment lengths (See example cDNA profile in Fig. 2.).
8) Dilute cDNA
Dilute cDNA in water based on average concentration from Qubit measurements.
Target concentration is 150pg/µl per well.
9) Tn5 digestion
Tn5 is produced from psfTn5 (Addgene #79107), purified to ~3mg/ml and assembled with Illumina Tn5 adapters (see oligos in Reagents) as in Picelli et al, 2014 (Ref. 1).
Prepare Tn5 master mix:
NOTE: TAPS-PEG Buffer contains PEG, which is viscous. Buffer should equilibrate to room temperature before use to allow proper mixing.
Reagent - Reagent volume
Nuclease free H2O - 1.05µl
TAPS-PEG (50mM TAPS, 25mM MgCl2, 40% PEG-8000) - 0.5µl
psfTn5 (loaded with 50µM MEDS-A/B) - 0.25µl
Total - 1.8µl
Dispense 1.8 µl per well in a new plate (tagmentation plate).
Add 0.7 µl cDNA (normalized to 150pg/µl) to tagmentation plate
Cover with new lid, mix well by vortexing plate and spin down.
Incubate in thermocycler at 55°C for 10 min.
Remove immediately and stop the reaction by adding 1µl of 0.1% SDS per well.
Vortex, spin down and incubate 7 min. at Room temperature
10) cDNA library PCR and barcoding
Make PCR master-mix:
Reagents - Reagent volume:
H2O - 13.25µl
5x buffer - 5µl
dNTPs - 0.75µl
KAPA - 0.5µl
Total - 19.50µl
Dispense 19.5 µl per well to tagmentation plate (containing 3.5 µl sample after step 9)
Add primers/barcodes:
Dispense 2 µl per well (from 384-well index plates, with 3.75µM/each forward/reverse primers; see oligos in materials).
Total reaction volume is 25 µl (3.5 µl sample + 19.5 µl PCR mix and 2 µl primers).
Cover, vortex and spin down. Incubate in thermocycler as follows:
1 cycle:
72ºC 3 min.
95ºC 30 sec
12 cycles:
95ºC 15 sec
67ºC 30 sec
72ºC 45 sec
1 cycle:
72ºC 4 min
4ºC hold
Following barcoding, pool 2.5µl from each well to an 1.5ml Eppendorf tube.
11) Library cleanup
i. Add 0.9x pooled library volume of SPRI-bead solution. Incubate for 5 min. at room temperature.
ii. Place on magnetic rack for 3 min.
iii. Remove supernatant without disturbing magnetic beads.
iv. Add at least 1 mL 80% EtOH (fresh). Incubate for 30 sec.
v. Remove supernatant.
vi. Repeat EtOH wash.
vii. Air dry for 10 - 15 min.
viii. Re-suspend beads thoroughly in 100 µl EB or TE buffer.
ix. Place eppendorf on magnetic rack for 3 min.
x. Transfer supernatant to new 1.5ml Eppendorf tube.
xi. Repeat cleanup (from step 1-7) and elute in 30µl EB or TE buffer.
xii. (Optional) Place eppendorf on magnetic rack for 3 min. and transfer supernatant to new tube.
12) Library quality control
Use Agilent HS 5000 DNA chips (or equivalent) to profile cDNA fragment lengths (See example cDNA profile in Fig. 3).
13) Sequencing
Pool libraries and sequence on Illumina sequencing platform of choice according to manufacturers instructions (For data generated in CIM-seq manuscript barcoded libraries were subjected to 75 bp single read sequencing on the Illumina NextSeq 550 instrument).
B) Feature selection and clustering (Seurat)
1. Data alignment and pre-processing
A series of pre-processing steps must be performed in order to generate a counts file:
i. Trim reads, remove adapter sequences and align RNAseq data to reference genome using STAR (Ref. 2)
ii. Remove duplicate reads using Picard (Ref. 3)
iii. Generate transcript counts file using HTSeq (Ref. 4)
2. Data preparation for CIM-seq
Once a counts file has been generated the data can be analyzed. CIM-seq requires five arguments in order to run:
i. The raw counts data with gene IDs as row names and sample IDs as column names.
ii. The ERCC spike-in counts data with gene IDs as row names and sample IDs as column names (Well based CIM-seq only).
iii. The dimensionality reduced representation of the data.
iv. A classification for each of the individual singlets.
v. A set of marker genes to be used for deconvolution
In order to generate the last three of these as well as removing low quality cells from the data set we recommend using the Seurat package in R (Ref. 5), as CIM-seq is implemented in R as well. A number of tutorials for Seurat can be found on the Satijalab website:
https://satijalab.org/seurat/index.html
C) Multiplet deconvolution
CIM-Seq can be downloaded from:
https://github.com/EngeLab/CIMseq
Or installed directly in R using the devtools package and running the command:
devtools::install_github("EngeLab/CIMseq")
The CIM-Seq Vignette should be followed in order to get an understanding for how the method works and the available arguments:
https://github.com/EngeLab/CIMseq/tree/master/vignettes