AbSeq Protocol Using the Nano-Well Cartridge-Based Rhapsody Platform to Generate Protein and Transcript Expression Data on the Single-Cell Level

Summary By including oligonucleotide-labeled antibodies into high-throughput single-cell RNA-sequencing protocols, combined transcript and protein expression data can be acquired on the single-cell level. Here, we describe a protocol for the combined analysis of over 40 proteins and 400 genes on over 104 cells using the nano-well based Rhapsody platform. We also include a workflow for sample multiplexing, which uniquely identifies the initial source of cells (such as tissue type or donor) in the downstream analysis after upstream pooling. For complete information on the use and execution of this protocol, please refer to Mair et al. (2020).

there are multiple approaches to ensure antibodies used for enrichment do not interfere with oligo-conjugated AbSeq reagents, including choices of non-competing clones. These considerations and others are discussed in more detail in the section ''Limitations''. For additional background on the technology underlying surface protein detection using oligo-nucleotide-labeled antibodies, we refer the reader to the following publications (Peterson et al., 2017;Stoeckius et al., 2017).
Note: The sample multiplexing capability is currently limited to a maximum of 12 different samples.
3. Immediately place the PBMC vials into a 37 C water bath. 4. For each vial, remove the PBMCs from the water bath when a small ice pellet remains. 5. Slowly add 1 mL of warm complete media to the PBMCs in the cryovial in a dropwise manner. 6. Transfer the contents of the vial dropwise to a 15 mL conical tube with 10 mL of pre-warmed complete media. 7. Using 1 mL of warmed complete media, rinse the vial to ensure collection of all cells and add to the 15 mL conical tube. 8. Centrifuge cells at 250 3 g for 5 min. 9. Decant the supernatant. 10. Resuspend pellet for each sample in 5 mL of warm complete media. 11. Count cells using Trypan blue to ascertain cell viability. 12. Take an aliquot containing the desired number of cells for analysis from each sample (it is recommended to start with at least three times the cell number that should be later loaded on the cartridge) and transfer them to new 1.5 mL LoBind tubes. 13. Centrifuge the tubes at 400 3 g for 5 min and remove the supernatant. 14. Resuspend the cell pellets in 180 mL of Sample Buffer.
Optional: If enrichment of certain cell populations is desired, use a cell sorter following standard practices (Cossarizza et al., 2019) and for each target population collect at least three times the cell number that should be later loaded on the cartridge. Keep sorted cells on ice, wash them after sorting and then resuspend them in 180 mL of cold Sample Buffer.
Note: all subsequent protocol steps can also be used with immune cells isolated from solid tissue samples after sort enrichment, or from fresh peripheral blood samples. CRITICAL: BD Stain Buffer contains Sodium Azide. Contact with acidic solutions and metal compounds over time may form potentially explosive metal azides. Should any of this material be introduced into a sanitary sewer system, flush with copious amounts of water.  ler et al., 2018) in R and RStudio, SeqGeq (BD Biosciences) can be used to analyze data. A comprehensive review of analysis packages for single-cell analysis pipelines is provided on https://osca.bioconductor.org/ (Amezquita et al., 2020).

STEP-BY-STEP METHOD DETAILS
Preparing the Cartridge

Timing: 30 min
The cartridge is essential for partitioning of single cells. There are two key steps to preparing the cartridge before addition of cells: priming the cartridge and treating the surface. It is possible to run two cartridges in parallel using a single Rhapsody Express instrument.  Figure 1). e. Move the front slider to WASTE f. Insert the tip of the pipette perpendicular to the port, seal the pipette tip against the gasket, and then load the cartridge with 700 mL of 100% (absolute) ethyl alcohol using the Gilson P1200M pipette in Prime/Treat mode ( Figure 2).
CRITICAL: ethyl alcohol should be steadily flowing through the cartridge. If ethyl alcohol is not going into the cartridge, the amount of pressure on the gasket from the pipette needs to be modified to allow reagent flow.
g. Load the cartridge with 700 mL of air using the Gilson P1200M pipette in Prime/Treat mode. h. Load the cartridge with 700 mL of Cartridge Wash Buffer 1 (PN 650000060) with the Gilson P1200M pipette in Prime/Treat mode ( Figure 2). i. Leave the cartridge on the tray at 15 C-25 C for 1 min. 4. Treating the surface of the cartridge a. Load the cartridge with 700 mL of air using the Gilson P1200M pipette in Prime/Treat mode. b. Load the cartridge with 700 mL of Cartridge Wash Buffer 1(PN 650000060) using the Gilson P1200M pipette in Prime/Treat mode (Figure 2). c. Leave the cartridge on the tray at 15 C-25 C for 10 min.
Note: If using two cartridges with a single instrument for one experiment, start the priming/ treating of the second cartridge during this 10-min incubation and repeat the steps above to prepare the second cartridge. d. Load the cartridge with 700 mL of air using the Gilson P1200M pipette in Prime/Treat mode.   This protocol describes use of antibodies conjugated to oligonucleotides (Ab-Oligo) for sample multiplexing and surface protein profiling. Each Ab-Oligo contains an antibody-specific barcode, a poly(A) tail for bead capture and additional sequences for PCR amplification and library generation. There are two different methods to utilize multiplexing and surface protein profiling: co-labeling or sequential labeling. Co-labeling will save time, however, sequential labeling is more economical and can additionally reduce batch effects resulting from inconsistent staining. The following steps describe sequential Ab-Oligo labeling of 20,000-1 million cells. Up to 100 antibodies can be pooled together per staining reaction. If no sample multiplexing is required, proceed directly to step 8 (Fc receptor blocking).
5. Use the cell suspensions derived from different samples (or from sorted populations) from the ''Defrosting cryopreserved Peripheral Blood Mononuclear Cells (PBMCs)'' section, and make sure that each sample is resuspended in 180 mL of Cell staining Buffer. 6. Labeling samples with sample multiplexing antibodies a. Prepare a spreadsheet listing which sample is going to be labeled with which Sample Tag (1-12). b. Quick-spin Sample Tag tubes to collect the contents at the bottom. c. For each sample, transfer 180 mL cell suspension to the corresponding Sample Tag tube.
Pipette-mix. d. Incubate at room temperature (15 C-25 C) for 20 min. e. Transfer each labeled cell suspension to a 1.5 mL LoBind tube. f. Add 1 mL Cell Staining Buffer to labeled cells and pipette-mix. g. Centrifuge each tube at 400 3 g for 5 min. h. Pipette off the supernatant taking care to not disturb the cell pellet. Note: Pipetting off supernatant as opposed to decanting will reduce cell loss during washes. However, it is critical to wash the cells well (i.e., leave as little supernatant as possible).
i. Repeat steps 6f-6h for a total of three washes j. Resuspend the cell pellet cells in 100 mL cold Cell staining buffer 7. Pooling the samples after labeling with multiplexing antibodies Optional: if the individual cell samples were obtained from different sources and the cell concentration is unknown, count the cells from step 6j using Trypan blue to determine the mixing ratio as required.
a. Pool all the individual samples from 6j into a new 1.5 mL LoBind tube. b. Add additional Sample Buffer to bring the total volume to 1.5 mL. c. Centrifuge the tube at 400 3 g for 5 min. d. Pipette off the supernatant taking care to not disturb the cell pellet. e. Resuspend the cells in 100 mL cold staining buffer.
Note: Samples containing myeloid and B cells should be treated with Fc-blocking reagent prior to Ab-Oligo staining (Andersen et al., 2016 Master mix can be scaled based on number of samples.
c. Pipette-mix the 23 AbSeq labeling master mix, and place back on ice.
Note: The final working concentrations for the Ab-Oligos are between 0.1 mg and 1 mg per stain depending on the antibody clone and have been optimized by the manufacturer. For an example of final working concentrations on a select set of Ab-Oligos, see Supplemental Table 4 in (Mair et al., 2020).
10. Labeling samples with AbSeq Ab-oligos a. In a new 1.5 mL LoBind tube, combine 100 mL pooled cell suspension labeled with Sample Tags (obtained from step 7e or 8h) and 100 mL 23 AbSeq labeling master mix. Pipette-mix. b. Incubate on Ice for 30 min. 11. Washing Labeled Cells a. Add 1 mL Cell staining Buffer to labeled cells and pipette-mix. b. Centrifuge the tube at 400 3 g for 5 min. c. Pipette off the supernatant taking care to not disturb the cell pellet. d. Repeat steps 11a-c for a total of 3 washes. e. Resuspend pellet in 100 mL cold Sample Buffer and use 10 mL to perform a Trypan blue viability staining. Ideally, if planning to load a full cartridge, the cell count should be >300,000 cells/mL with a viability >90%.

CRITICAL: Cells must be resuspended in Sample Buffer (and not Staining Buffer)
Note: Sufficient post-labeling washes are important for reducing noise that comes from residual unbound antibodies being captured onto 3' capture beads during single-cell capture. However, some cell loss occurs with each additional wash. Users can choose to perform more or fewer washes depending on the abundance of their sample. If the user has an excess of cells for the experiment, it is recommended that the stain and wash steps occur in a 5 mL polystyrene tube. This will allow an increase in wash volume to 3 mL instead of 1 mL.

Loading of Cells onto the Cartridge and Retrieval of Cell Capture Beads Containing mRNA and Feature Barcodes
Timing: 1 h This protocol describes the steps to isolate and capture mRNA and feature barcodes from partitioned cells on the Rhapsody cartridge. Single-cell partitioning is achieved by loading cells at densities that are low enough to achieve a single cell per well distribution in the cartridge. Each cartridge contains >200,000 nanowells, thus the likelihood of having more than one cell in a single well depends on the cell density (see table below). There are different approaches to identify multiplets based on factors such as number of genes or unique molecular identifiers present compared to the rest of the sample, as well as computational approaches (Amezquita et al., 2020). Identifying multiplets can be significantly improved if several samples are multiplexed prior to loading on the cartridge because any sample with two or more Sample Tags is easily identified as a multiplet (Stoeckius et al., 2018).
Refer to Table 3 to determine an acceptable multiplet rate for the number of captured cells on retrieved Cell Capture Beads: An approximation of the percentage of multiplets that will occur within a cartridge based on the number of cells loaded. Because there are a fixed number of nanowells in a cartridge, the likelihood of multiplets increases with the number of loaded cells. As discussed in the main text, multiplets can only efficiently be removed when using sample multiplexing.
12. According to the number of cells counted in step 11e, take the appropriate volume containing the targeted cell number + 30% (i.e., if planning to capture 20,000 cells per cartridge, take 26,000 cells. If planning to load two cartridges in parallel take 52,000 cells).
Note: 30% is an approximate number that should allow a more accurate prediction of the number of cells collected after. This number takes into account cell loss from both washes and cartridge loading (only 575 mL of the 650 mL single-cell suspension will be loaded onto the Rhapsody Cartridge).
13. Add Sample Buffer to bring the volume of the cell suspension to 650 mL (or 1,300 mL if planning to load two cartridges in parallel) 14. Load the cells onto the cartridge.
a. Load the cartridge on the tray with 700 mL of air using the Gilson P1200M pipette in Prime/ Treat mode. b. Change the mode of the Gilson P1200M pipette to Cell Load. c. With a manual pipette, gently pipet the cell suspension up and down to mix. d. On the Gilson P1200M, press the pipette button once to aspirate 40 mL of air, immerse the pipette tip in cell suspension, and then press the button again to aspirate 575 mL of cold cell suspension ( Figure 2). e. Insert the tip of the pipette perpendicular to the port, seal the pipette tip against the gasket, and then press the button a third time to dispense 615 mL of air and cells. f. Incubate at room temperature (15 C-25 C) for 15 min. During the 15-min incubation, prepare Cell Capture Beads.
CRITICAL: If planning to load two cartridges at the same time using only one Rhapsody Express instrument, load cartridge #2 at the 14-min mark of step 14f by repeating steps 14a-f. This is essential to ensure you have correct timing for subsequent steps.  18. During 2-min lysis step:

Preparing Cell Capture Beads
a. Ensure that a 5 mL LoBind Tube (Eppendorf cat. no. 0030108310) was inserted into the drawer for bead retrieval. b. Set the mode on the Gilson P5000M pipette to Retrieval. c. Move the front slider to BEADS. CRITICAL: If the slider is not set to BEADS, then sample will go into the waste and the experiment is lost.   g. Move the front slider to OPEN, and then remove and cap the 5 mL LoBind Tube. h. Uncap the tube and place it on the large magnetic separation stand fitted with the 15 mL tube adapter for 1 min. Proceed immediately to Washing Cell Capture Beads. i. Appropriately dispose of cartridge, waste collection container, and Lysis Buffer with DTT. 20. Washing Cell Capture Beads a. After 1-min incubation leaving the 5 mL tube containing retrieved Cell Capture Beads on large magnet, remove all but 1 mL of supernatant without disturbing beads ( Figure 6). b. Remove tube from magnet. Gently pipette-mix beads, and transfer to a new 1.5 mL LoBind tube. c. If there are still beads left in the 5 mL tube, add 0.5 mL Lysis Buffer with DTT, rinse 5 mL tube, and transfer to 1.5 mL LoBind tube from step 20b. d. Place tube on magnet for % 2 min and remove supernatant.
CRITICAL: Avoid leaving Lysis Buffer or bubbles in tube. Lysis Buffer might cause the reverse transcription reaction to fail.
e. Remove tube from magnet and pipette 1 mL of cold Bead Wash Buffer into tube. Pipet mix. f. Place tube on 1.5 mL tube magnet for %2 min and remove supernatant. g. Remove tube from magnet, and pipette 1 mL cold Bead wash Buffer into tube. Pipette-mix, and place on ice.
CRITICAL: Start reverse transcription %30 min after washing retrieve Cell Capture Beads with Bead Wash Buffer. If a second cartridge is used, then repeat steps 17a-20g with the second cartridge within this time.
Performing Reverse Transcription and Exonuclease I Treatment on the Cell Capture Beads

Timing: 90 min
This protocol describes the steps to link cell barcodes to mRNA and feature barcodes by reverse transcription, which yields a stable sample and preserving single-cell information for downstream library generation steps. If two cartridges were used in parallel, the two sets of Cell Capture Beads should be kept separate and be treated as two independent libraries.
21. Performing reverse transcription a. In the pre-amplification workspace, in a new 1.5 mL LoBind Tube that is on ice, pipette the components in the following order to prepare the cDNA mix (Table 4): Add reagents sequencially, as listed in table, into a 1.5 mL LoBind tube on ice. Volumes can be further scaled up if more than 2 cDNA libraries need to be produced.  c. Place the tube of washed beads on the 1.5 mL tube magnet for %2 min, and then carefully remove and appropriately discard the supernatant without disturbing the beads while leaving the tube on the magnet. d. Use a low retention tip to pipette 200 mL of the cDNA mix to resuspend the beads. Gently mix the suspension by pipette only. Do not vortex. e. Transfer the bead suspension to a new 1.5 mL LoBind Tube. f. Ensure that the SmartBlock Thermoblock 1.5 mL or equivalent is installed on the thermomixer. g. Incubate the suspension on the thermomixer at 1,200 rpm and 37 C for 20 min. h. After incubation, put the tube on ice. 22. Treating the Cell Capture Beads with Exonuclease I a. In the pre-amplification workspace, prepare the Exonuclease I mix in a new 1.5 mL LoBind Tube that is on ice by adding the components in the following order (Table 5): Into a 1.5 mL LoBind tube on ice, add reagents sequencially, as listed in table. Exonuclease I mix can be scaled up to accommodate more than two samples.
b. Gently vortex and centrifuge the mix, and then put it back on ice. c. Place the tube of beads with cDNA mix on the 1.5 mL tube magnet for %2 min, and then carefully remove and appropriately discard the supernatant without disturbing the beads and while leaving the tube on the magnet. d. Remove the tube from the magnet, and then use a low retention tip to pipette 200 mL of Exonuclease I mix to the tube, gently resuspend the beads by pipette only. Do not vortex. e. Incubate the suspension on the thermomixer at 1,200 rpm and on the thermomixer at 1,200 rpm and 37 C for 30 min. 23. Inactivating Exonuclease I a. Transfer the bead suspension with Exonuclease I to the thermomixer or heat block in the preamplification workspace at 80 C (no shaking) for 20 min.
Note: If the thermomixer is the same as that used for the 37 C step put the samples on ice until that temperature is reached rather than leaving the tubes in the thermomixer as the temperature ramps up.
b. Put the bead suspension on ice for $1 min. c. Place the tube on the 1.5 mL tube magnet until the solution is clear (%1 min). d. Carefully remove and appropriately discard the supernatant without disturbing the beads while leaving the tube on the magnet. e. Remove the tube from the magnet, and with a low retention tip, pipette 200 mL of cold Bead Resuspension Buffer to gently resuspend the beads. Do not vortex.
Pause Point: The Exonuclease I-treated beads can be stored at 2 C-8 C for %3 months. Note: Only remove enzymes from À20 C when in use.

Figure 7. Library Preparation Workflow
Overview of each PCR required for library preparation. After PCR1, DNA double-sided size selection will split the transcripts based on amplicon size into mRNA or Sample Tag/AbSeq libraries. PCR2 is then preformed to further enrich both mRNA and Sample Tag libraries. Finally, each library is indexed for sequencing.

OPEN ACCESS
25. Perform PCR1 a. In pre-amplification workspace, pipette reagents into a new 1.5 mL LoBind Tube on ice (Table 6): Combine all components into a 1.5 mL LoBind tube on ice. If more than 2 PCR reactions are needed, volumes can be scaled accordingly.
b. Gently vortex mix, briefly centrifuge, and place back on ice. c. Place tube of Exonuclease I-treated beads in Bead Resuspension Buffer (Cat. No. 650000066) on 1.5 mL magnet for <2 min. Remove supernatant. d. Remove tube from magnet and resuspend beads in 200 mL PCR1 reaction mix. Do not vortex. e. Ensuring that the beads are fully resuspended, pipette 50 mL PCR1 reaction mix with beads into each of four 0.2 mL PCR tubes. f. Bring reaction mix to the post-amplification workspace. g. Program the thermal cycler (Table 7): Depending on the number of cells targeted, a different number of cycles should be used to ensure sufficient quantity of library and limit PCR bias. This table is based on resting PBMCs, and should be adjusted accordingly for different cell types. Note that the targeted cell number is lower than the actual number of cells loaded on the cartridge (as calculated in step 12).
h. Ramp heated lid and heat block of post-amplification thermal cycler to R95 C by starting the thermal cycler program and then pausing it. i. For each 0.2 mL PCR tube, gently pipette-mix, immediately place tube in thermal cycler, and unpause the thermal cycler program.
Program a thermocycler to run the above steps. See Table 8 for suggested number of cycles. Do not use fast cycling mode. CRITICAL: The thermocycler should be at 95 C when the tubes are added to ensure amplification of on-bead molecules.
Pause Point: Once the PCR1 cycle is complete, the reaction can be stored at 4 C. Purification of PCR1 must occur %24 h after PCR1.
j. After PCR, briefly centrifuge tubes. k. Pipette-mix and combine the four reactions into a new 1.5 mL LoBind Tube. l. Place the 1.5 mL tube on magnet for 2 min and carefully pipette the supernatant (PCR1 products) into a new 1.5 mL LoBind Tube without disturbing the beads.
Note: (Optional) Remove tube with the Cell Capture Beads from magnet, and pipette 200 mL cold Bead Resuspension Buffer into tube. Pipette-mix. Do not vortex. Store beads at 2 C-8 C in post-amplification workspace. Beads can be sub-sampled or undergo multiple rounds of amplification with different primer panels.
26. Perform double-sided SPRISelect bead purification to separate the shorter AbSeq and Sample Tag PCR1 products ($170 bp) from the longer mRNA targeted PCR1 products (350-800 bp). a. Prepare 5 mL fresh 80% (v/v) ethyl alcohol by combining 4.0 mL absolute ethyl alcohol, molecular biology grade (major supplier) with 1.0 mL nuclease-free water (major supplier). Vortex tube for 10 s. b. Vortex SPRISelect Reagent at high speed 1 min until beads are fully resuspended. c. Pipette 140 mL SPRISelect beads into the tube with 200 mL PCR1 products (step 25l). Pipettemix 10 times.
CRITICAL: SPRISelect beads should be pipetted very carefully. To ensure appropriate size selection, it is essential that only the exact amount of SPRISelect beads is added to PCR1 products. Do not immerse pipette tips with SPRISelect bead droplets on the outside into PCR1 products. Instead, place tip on side of tube and slowly expel all liquid, replace pipette tip, and pipette-mix.
d. Incubate at room temperature (15 C-25 C) for 5 min. e. Place 1.5 mL LoBind Tube on magnet for 5 min. f. Keeping tube on magnet, transfer the 400 mL supernatant (AbSeq PCR1 and Sample Tag products) to a new 1.5 mL tube without disturbing beads (mRNA targeted PCR1 products). g. Store the supernatant at room temperature (15 C-25 C) while purifying and eluting the mRNA targeted PCR1 products (27a-27h below), then purify the AbSeq and Sample Tag PCR1 products. 27. Purifying mRNA targeted PCR1 products a. Keeping tube on magnet, gently add 500 mL fresh 80% ethyl alcohol to the tube of SPRISelect beads bound with mRNA targeted PCR1 products and incubate 30 s. Remove supernatant. b. Repeat step 27a once for two washes. c. Keeping tube on magnet, use a small-volume pipette to remove residual supernatant from tube, and discard. d. Air-dry beads at room temperature (15 C-25 C) for 5 min. e. Remove tube from magnet and resuspend bead pellet in 30 mL of Elution Buffer (Cat. No. 91-1084). Vigorously pipette-mix until beads are uniformly dispersed. Small clumps do not affect performance. f. Incubate at room temperature (15 C-25 C) for 2 min, and briefly centrifuge. g. Place tube on magnet until solution is clear, usually %30 s. h. Pipette the eluate ($30 mL) into a new 1.5 mL LoBind Tube (purified mRNA targeted PCR1 products).
Pause Point: Store at 2 C-8 C before proceeding in %24 h or at -25 C to -15 C for % 6 months.
28. Purifying combined AbSeq/Sample Tag PCR1 products a. Pipette 100 mL SPRISelect beads into the tube with 400 mL AbSeq/Sample Tag PCR1 products from step 26g above. Pipette-mix 10 times. b. Incubate at room temperature (15 C-25 C) for 5 min. c. Place on magnet for 5 min. d. Keeping tube on magnet, remove supernatant. e. Keeping tube on magnet, gently add 500 mL fresh 80% ethyl alcohol, and incubate 30 s. Remove supernatant. f. Repeat step 28e once for two washes. g. Keeping tube on magnet, use a small-volume pipette to remove residual supernatant from tube, and discard. h. Air-dry beads at room temperature (15 C-25 C) for 5 min.
i. Remove tube from magnet and resuspend bead pellet in 30 mL Elution Buffer (Cat. No. 91-1084). Vigorously pipette-mix until beads are uniformly dispersed. Small clumps do not affect performance. j. Incubate at room temperature (15 C-25 C) for 2 min, and briefly centrifuge. k. Place tube on magnet until solution is clear, usually %30 s. l. Pipette the eluate ($30 mL) into a new 1.5 mL LoBind Tube (purified AbSeq/Sample Tag PCR1 products).
Pause Point: Store at 2 C-8 C before proceeding in %24 h or at -25 C to -15 C for % 6 months. 29. Quantify AbSeq/Sample Tag PCR1 products a. Dilute an aliquot ($2 mL) 1:1 with nuclease-free water, and run on the Agilent TapeStation with the High Sensitivity D5000 ScreenTape. b. Measure the yield of the largest peak of the AbSeq/Sample Tag PCR1 products ($170 bp, Figure 8).
CRITICAL: Do not use fast cycling mode c. Dilute an aliquot of AbSeq/Sample Tag PCR1 products to 0.1-1.1 ng/mL with Elution Buffer (Cat. No. 91-1084) before index PCR of AbSeq PCR1 products. Use undiluted AbSeq/Sample Tag PCR1 products for Sample Tag PCR2 amplification. 30. Perform PCR2 of targeted mRNA (Table 9) and Sample Tag products (Table 10). The AbSeq PCR1 products do not require additional amplification beyond index PCR.
a. In pre-amplification workspace, pipette reagents into a new 1.5 mL LoBind Tube on ice: This reaction will further amplify the mRNA products. Add all components to a 1.5 mL LoBind tube on ice.
This reaction will amplify the Sample Tag products from PCR1. Combine all components from this table into a 1.5 mL LoBind tube on ice.
b. Gently vortex mix, briefly centrifuge, and place back on ice. c. Bring PCR2 mixes into post-amplification workspace. d. In two separate, new 0.2 mL PCR tubes: i. mRNA targeted PCR1 products: Pipette 5.0 mL products into 45.0 mL mRNA targeted PCR2 reaction mix. ii. Sample Tag PCR1 products: Pipette 5.0 mL products into 45.0 mL Sample Tag PCR2 reaction mix.  (Table 11).
Program a thermocycler to this program without using fast cycling mode.
Pause Point: PCR2 products can be stored at 2 C-8 C for %24 h or stored at À25 C to À15 C for %6 months. 31. Purify targeted mRNA and Sample Tag PCR2 products a. Vortex SPRISelect beads at high speed 1 min until beads are fully resuspended. b. Briefly centrifuge PCR2 products. c. To 50.0 mL PCR2 products pipette: i. mRNA targeted PCR2 products: 40 mL SPRISelect beads.
ii. Sample Tag PCR2 products: 60 mL SPRISelect beads. d. Pipette-mix 10 times, and incubate at room temperature (15 C-25 C) for 5 min. e. Place tube on strip tube magnet for 3 min. Remove supernatant. f. Keeping tube on magnet, gently add 200 mL fresh 80% ethyl alcohol into tube, and incubate 30 s. Remove supernatant. g. Repeat step 31f once for two washes. h. Keeping tube on magnet, use a small-volume pipette to remove residual supernatant from tube and discard. i. Air-dry beads at room temperature (15 C-25 C) for 3 min. j. Remove tubes from magnet, and resuspend bead pellet in 30 mL Elution Buffer (Cat. No. 91-1084). Pipette-mix until beads are fully resuspended. k. Incubate at room temperature (15 C-25 C) for 2 min, and briefly centrifuge. l. Place tubes on magnet until solution is clear, usually %30 s. m. Pipette entire eluate ($30 mL) of each sample into two separate new 1.5 mL LoBind Tubes (purified mRNA targeted PCR2 and Sample Tag PCR2 products).
Pause Point: Store at 2 C-8 C before proceeding on the same day or at -25 C to -15 C for %6 months.
n. Estimate the concentration of each sample by quantifying 2 mL of the PCR2 products with a Qubitä Fluorometer using the Qubit dsDNA HS Assay Kit. o. Dilute an aliquot of the products with Elution Buffer (Cat. No. 91-1084): i. mRNA targeted PCR2 products: 0.2-2.7 ng/mL. ii. Sample Tag PCR2 products: 0.1-1.1 ng/mL. 32. Perform index PCR to prepare final libraries a. In pre-amplification workspace, prepare the three libraries + 20% overages of the final amplification mix for each of the three products. Pipette reagents into a new 1.5 mL LoBind Tube on ice (Table 12): Combine all components in a 1.5 mL LoBind Tube on ice. Utilize a different reverse primer for each sample that will be run on the same sequencing flow cell. The mRNA, AbSeq, and Sample Tag libraries from the same cartridge should use the same reverse primer.
b. Gently vortex mix, briefly centrifuge, and place back on ice. c. Bring index PCR mixes to post-amplification workspace. d. In three separate, new 0.2 mL PCR tubes: i. mRNA targeted PCR2 products: Pipette 3.0 mL of 0.2-2.7 ng/mL products into 47.0 mL index PCR mix. ii. Sample Tag PCR2 products: Pipette 3.0 mL of 0.1-1.1 ng/mL products into 47.0 mL index PCR mix. iii. AbSeq PCR1 products: Pipette 3.0 mL of 0.1-1.1 ng/mL products into 47.0 mL index PCR mix. e. Gently vortex and briefly centrifuge. f. Program the thermal cycler. (Tables 13 and 14) Program a thermocycler to run this program. Refer to Table 14 for cycle numbers.
This table is a suggestion for the number of index cycles to use depending on the concentration of each library. It is possible you may need to adjust the number of cycles per experiment.

OPEN ACCESS
Pause Point: The Index PCR product can be stored at 2 C-8 C for %24 h, or at À25 C to À15 C for %6 months.
33. Purify index PCR products a. Vortex SPRISelect beads at high speed 1 min until beads are fully resuspended. b. Briefly centrifuge index PCR products. c. To 50.0 mL of each of the individual index PCR products pipette: i. mRNA targeted library: 35 mL SPRISelect beads.
ii. AbSeq and Sample Tag libraries: 40 mL SPRISelect beads. d. Pipette-mix 10 times. Incubate at room temperature (15 C-25 C) for 5 min. e. Place each tube on strip tube magnet for 3 min. Remove supernatant. f. Keeping tube on magnet, for each tube, gently add 200 mL fresh 80% ethyl alcohol into tube and incubate 30 s. Remove supernatant. g. Repeat step 33f for a second wash h. Keeping tubes on magnet, use a small-volume pipette to remove residual supernatant from tube, and discard. i. Air-dry beads at room temperature (15 C-25 C) for 3 min. j. Remove tubes from magnet and resuspend each bead pellet in 30 mL Elution Buffer (Cat. No. 91-1084). Pipette-mix until beads are fully resuspended. k. Incubate at room temperature (15 C-25 C) for 2 min, and briefly centrifuge. l. Place tubes on magnet until solution is clear, usually %30 s. m. For each tube, pipette entire eluates ($30 mL) into three separate new 1.5 mL LoBind Tubes (final sequencing libraries). 34. Perform quality control on the final sequencing libraries a. Estimate the concentration of each sample by quantifying 2 mL of the final sequencing library with a Qubit Fluorometer using the Qubit dsDNA HS Kit, following manufacture's protocol, to obtain an approximate concentration of PCR products to dilute for quantification on an Agilent 4200 TapeStation. Follow the manufacturer's instructions. The expected concentration of the libraries is >1.5 ng/mL. b. Measure the average fragment size of the mRNA targeted library by using the Agilent TapeStation with the High Sensitivity D5000 tape following the manufacturer's instructions. c. The final mRNA targeted library should show a fragment distribution that depends on the panel used (Figure 9). The expected size of Sample Tag index PCR product is 290 bp (Figure 10). You might observe a smaller peak of $270 bp, which corresponds to AbSeq product. The expected size of AbSeq index PCR products is $270 bp ( Figure 11). Figure 9. Example of Final mRNA Library TapeStation Trace The major peak length will depend on the PCR panels used. If you detect some transcripts with longer base pair lengths, this is typically not an issue as shorter transcripts amplify more efficiently than longer products.

CRITICAL: Do not use fast cycling mode:
Note: when preparing the libraries from two independent cartridges, prepare two separate master mixes with unique reverse primers. For information on ordering additional indices beyond the four indices included with the reagent kit contact scomix@bdscomix.bd.com.   Table 15).
Suggested flowcell loading concentrations for libraries and PhiX for different Illumina systems.
b. Next determine how many reads each sample requires. To do this, multiply the target cell number by the desired number of reads per cell. For example, if you wanted to sequence sample one mRNA library at 5,000 reads per cell, and had a target number of 10,000 cells, the number of reads for the sample would be 5,000 3 10,000 = 5,000,000 reads. c. Determine the fraction of the reads will go toward each sample. This can be done by dividing by each sample by the total reads needed across all samples. If you wanted to multiplex the Sample Tag, mRNA, and AbSeq libraries from a sample and each has 5,000,000 reads, 50,000,000 reads, and 150,000,000 reads respectively, the total reads is 205,000,000. The fraction of reads that will go to the Sample Tag is 0.02. d. To determine the volume (mL) of each sample diluted to the flowcell loading concentration needed to pool for the final multiplexed library, multiply the fraction of reads by the total volume of the library needed. For example, to calculate the volume of Sample Tag (from step 35c) you need in your final library of 150 mL, multiply 0.02 by 150 to determine you will need 3.66 mL.
Note: If the volume of library requires is prohibitively small a higher volume can be used for the pooling. For example, if only 5 mL is needed, a 100 mL pool can be made, and 5 mL of that pool can be taken into sequencing.
e. Pipette each required amount of sample into a LoBind 1.5 mL tube. 36. Sequence the libraries a. Recommended sequencing depths are as follows: i. mRNA: 5,000-10,000 reads per cell ii. Sample Tag: 600-1,000 reads per cell iii. AbSeq: 400-1,000 reads per cell per antibody Note: The recommended sequencing depths are provided as a starting point, since the number of reads that should be targeted is not a fixed value, but rather depends on several factors. First, the number of reads needed per cell for each mRNA library will depend on how many primers are included in your panel and the abundance of the transcripts of interest. If the goal is to capture rare transcripts as much as possible, a higher read per cell sequencing target may be beneficial. Similarly, the targeted reads per cell for each AbSeq library will depend on the makeup of the antibody panel. Proteins that are expressed at high levels will use proportionally more reads, which can leave other antibodies in the panel poorly resolved, particularly if they target proteins that are expressed at low levels.

EXPECTED OUTCOMES
Type of data that are generated: The protocol presented here details sample and library preparation for the detection of targeted transcript and surface protein expression at the single-cell level. After sequencing and pre-processing of the Fastq files, the final output will essentially yield a large and sparse data matrix containing the molecule counts for mRNA and the proteins for each cell (MolsPerCell.csv).
For the analysis of these complex data sets, there are different options available. Users that prefer a graphical user interface can use commercial software packages such as SeqGeq or BioTuring (among others). If more versatile options are required it is recommended to use the R environment and dedicated packages for single-cell analysis which are commonly available on Bioconductor. A full description of this process and all available analysis approaches is beyond the scope of this protocol and we refer the reader to recent reviews which also include excellent online tutorials and links to additional resources (Amezquita et al., 2020;Luecken and Theis, 2019).
The exact number of cells that are called in the sequencing data will be less than the number of cells loaded on the cartridge. The number depends on the sample type, accuracy of cell counts and viability of the cells loaded. For a typical experiment with human leukocytes, after quality control filtering and removal of multiplets, users can expect to recover data for about 60% of the total cells loaded.
The final number of gene targets varies depending on the design of the target transcriptomic assay but can be up to 499 target genes. We reported that with an assay targeting 492 genes, 2,000-4,000 reads per cell from the transcriptomic portion of the library delivered sufficient resolution 1 . Like the transcriptomic analysis, the number of proteins targeted can be customized for specific experimental aims. We previously reported that with an assay targeting 41 proteins, read depths of 200-400 reads/antibody/cell from the protein portion of the library results in sufficient resolution 1 .

LIMITATIONS Limitations to Assessing Protein Expression
Antibody-based probes are powerful tools for detection of specific protein targets. However, meaningful interpretation of sequencing-based protein measurements requires rigorous validation. In some instances we observed that the same antibody clone can yield different expression patterns in sequencing-based analysis relative to conventional cytometry (Mair et al., 2020). Furthermore, the sequencing read depth in context of the chosen antibody panel is of particular importance here. Highly expressed proteins (e.g., lineage markers such as CD4, CD8, HLA-DR) will use a significant portion of the sequencing reads, which may make it difficult to resolve less abundant proteins (such as IL-7Ra) if the chosen read depth is too low. While there is no general consensus yet for a minimum read depth in all experimental settings, it is recommended to calculate with at least 200-400 reads per antibody per cell as a rule of thumb (Mair et al., 2020).

Limitations to Assessing Transcript Expression
Our previous comparison between whole transcriptome and targeted transcript analysis demonstrated that targeted analysis can be highly sensitive in detection of low-abundance transcripts, but some genes may be under-represented compared to whole transcriptome analysis (and vice versa). This may result from different amplification efficiencies between multiplexed targeted primers used in the Rhapsody platform and the template-switch process in other WTA platforms. A lack of a signal for a specific transcript (by either WTA or targeted transcriptomics) cannot ll OPEN ACCESS STAR Protocols 1, 100092, September 18, 2020 necessarily be interpreted as absence of transcript expression. Further, transcripts that contain internal poly A stretches can also artificially inflate the number of captured transcript molecules with WTA that would otherwise correctly not be captured in a 3' end targeted approach. Differences in poly A site expression (location and total number) due to tissue types and cell states (i.e., activated versus resting) can cause inherent heterogeneity in abundance and identity of observed transcripts. Because of these issues one possible strategy is to first use WTA on the sample of interest to inform locations of poly A sites to include in the primer design of targeted mRNA panels to better correlate these two types of data. Finally, it is important to remember in this context that the dynamic range of expression for proteins spans about 6-7 orders of magnitude, whereas transcript copy numbers span about two orders of magnitude (Azimifar et al., 2014;Schwanhausser et al., 2011), with a mean copy number of $4 copies per cell, which are inherently difficult to detect with any single-cell analysis approach.

Limitations to the Number of Cells that Can Be Analyzed and Enrichment Strategies
If the experimental goal is an exploratory snapshot of all immune cells, one may start with a bulk immune cell population (e.g., human peripheral blood lymphocytes). However, if a specific and possibly rare cell subset is of interest, then enriching for these cells may be necessary to obtain meaningful data. For example, human dendritic cell subsets might represent only 10-100 cells in a typical PBMC sample of 20,000 cells, thus requiring enrichment for meaningful analysis.
If enrichment by cell sorting (FACS) is performed, it is important to ensure that the used fluorochrome-conjugated antibodies do not overlap or interfere with the antibody clones used in the downstream oligo-antibody assay. This can be tested e.g., by prior co-staining of different antibody clones by flow cytometry and selecting only compatible clones. As an alternative, if the same clones have to be used, it is possible to evaluate the cell staining ratios of fluorochrome-conjugated antibodies with AbSeq antibodies coupled with complementary sequence-conjugated fluorochrome (Flow Proxy) to determine an optimal ratio for cell sorting while maintaining AbSeq staining. In these scenarios, fluorochrome-conjugated Abs can be stained simultaneously with oligo-conjugated Ab-Seq reagents saving researchers time and minimize handling and potential perturbations to cells.
Importantly, if magnetic enrichment protocols are used, positive selection magnetic bead enrichment should be avoided while negative selection/depletion is preferred. Ferrous particles that remain on cells from enrichment may interfere with downstream processing in the nano-wellcartridge.

TROUBLESHOOTING Problem 1
Final library concentration lower than what is required for sequencing.

Potential Solution
If the final library has too low of a concentration, the indexing PCR can be re-done with the PCR2 products using more than the recommended 6-8 cycles. The additional number of cycles will depend on how much more the final library concentration needs to increase. It is recommended that you start with just one or two extra cycles more to avoid as much PCR amplification bias as possible. Another option is to return to PCR1 products, or completely remake the library from the saved exonuclease treated beads.

Problem 2
Tape Station traces are showing peaks that are not very clean (see Figure 12).

Potential Solution
This is generally due to poor SPRISelect bead clean up. Ensure you are using exactly the indicated volume of beads and there are no SPRISelect bead droplets on your pipette tips. The SPRISelect ll OPEN ACCESS bead clean up can be repeated on a sample that has already been cleand up, using the same ratio of beads to sample as before, although some sample loss should be expected.

Problem 3
Visualization of AbSeq data in FlowJo does not show the expected expression patterns.

Potential Solution
One of the possible analysis options is to convert the molecule per cell CSV-file into the flow cytometry standard (FCS) file format by using R packages such as flowCore or premessa. This allows the visualization of bivariate plots in the popular software package FlowJo. However, the data ranges for RNA and protein expression are widely different, and thus it is important to adjust the transformation of these two molecule classes appropriately. One option is to choose the arcsinh transform in FlowJo and adjusting the co-factor as required to achieve bimodal expression for well-defined protein markers (e.g., CD3).

Problem 4
No signal obtained for oligo-antibody conjugate after prior enrichment with cell sorting.

Potential Solution
The reason for this can be that a competing fluorescent-labeled antibody clone was used for cell enrichment, preventing the subsequent binding of the oligo-antibody conjugate to its target. As discussed in more detail in the section ''Limitations'', if the same antibody targets are needed for enrichment as well as for readout during the multi-omic workflow, it is recommended to use different antibody clones. Experimenters need to perform prior co-staining experiments using these clones in a conventional flow cytometry experiment to test whether the clones are compatible (e.g., for targeting human CD3, the two clones OKT3 and SK7 can be used).

Problem 5
Poor resolution obtained for oligo-antibody signal.

Potential Solution
For some antibodies and experimental setups, the final concentration as pre-determined by the manufacturer might be suboptimal. In this case, it is recommended to titer the oligo-antibody conjugate on a conventional flow cytometer using an oligo-nucleotide coupled to fluorochromes (oligo-dT-fluorochrome).

RESOURCE AVAILABILITY Lead Contact
Further information and requests for resources and reagents that are not commercially available should be directed to and will be fulfilled by the Lead Contact, Martin Prlic (mprlic@fredhutch.org).

Materials Availability
This study did not generate new unique reagents.

Data and Code Availability
The published article includes all code generated during this study. Optional code for data processing post Seven Bridges (not covered in this protocol) are available at GitHub: https://github.com/ MairFlo/Targeted_transcriptomics