Droplet-Based 10x Genomics 3' scRNA-seq Capture and GEM Generation from Dissociated Mouse Spleen Single-Cell Suspensions

Three biological replicate mice per condition; one 10x channel per mouse (8,000-cell target). Cell concentration is standardized to 700-1,200 cells/µL by hemocytometer/automated count after viability staining. Samples are processed within 30 min of dissociation to limit stress-response gene induction. Genetic-demultiplexing or CellPlex hashing allows pooling of replicates to reduce batch effects. Loading volume is calculated from the 10x cell-recovery table; the operator is blinded to condition during loading and QC.

BSL-1/ABSL-1 for routine mouse spleen (BSL-2 if infected/transgenic-pathogen models). All animal work under an approved IACUC protocol; euthanize per approved method. Handle tissue in a biosafety cabinet, wear gloves and lab coat. Collagenase and ACK buffer are low hazard; 10x partitioning oil and recovery agents are mineral-oil/fluorinated reagents — avoid skin contact and follow SDS. Autoclave or incinerate animal tissue waste; sharps in approved containers.

Viability control: AOPI/trypan staining gating to ensure >85% viable cells loaded (dead cells inflate ambient RNA). A single-species loading control (or species-mixing experiment with human + mouse cells) estimates the true multiplet rate. No-cell GEM run (water in place of cells) checks for barcode/ambient contamination. A reference cell line (e.g., known PBMC or cell-line standard) run periodically benchmarks instrument and reagent performance. Empty-droplet/ambient profiles serve as an internal background control.

Recovered cells ~7,000-9,000 per channel at an estimated multiplet rate ~6%. cDNA Bioanalyzer trace shows a broad peak 400-9,000 bp. Post-sequencing (CellRanger): >85% valid barcodes, median genes/cell >2,000, median UMIs/cell >5,000, mitochondrial reads <10%, sequencing saturation 40-70%, and clearly separable T/B/myeloid clusters on UMAP. Library insert ~350-450 bp, yield 5-30 nM.

• Mouse spleen, freshly harvested into cold RPMI-1640
• Collagenase/DNase dissociation mix (or mechanical dissociation through 70 µm strainer)
• RBC lysis buffer (ammonium-chloride-potassium, ACK)
• Dead-cell removal kit (magnetic, annexin-based) or MACS dead-cell removal
• AOPI or trypan blue viability stain; PBS + 0.04% BSA (resuspension)
• 10x Chromium Next GEM Single Cell 3' Gel Beads kit (v3.1) + Chip G + partitioning oil
• Master mix: RT enzyme, template-switch oligo, cell-barcoded gel beads, UMIs
• Dynabeads/SPRI (AMPure XP) for cleanup
• cDNA amplification mix; fragmentation/end-repair/A-tailing module; adapter + Dual Index plate
• Qubit dsDNA HS; Bioanalyzer/TapeStation HS

To generate barcoded single-cell 3' gene-expression libraries from a viable single-cell suspension of dissociated mouse spleen via droplet partitioning (GEMs) on the 10x Chromium Controller, reverse transcription with cell-barcoded/UMI-tagged primers, cDNA amplification, and Illumina library construction. The goal is to recover ~8,000 high-quality single cells per channel with >85% viability, median >2,000 genes/cell, and ambient/empty-droplet contamination minimized.

1. Dissociate spleen mechanically through a 70 µm strainer in cold RPMI; pellet 300 x g, 5 min, 4°C.
2. RBC lyse with ACK buffer 3 min on ice; quench with PBS+BSA; pellet.
3. Filter through 40 µm strainer; perform dead-cell removal; resuspend in PBS+0.04% BSA.
4. Count with AOPI; require >85% viability; adjust to ~1,000 cells/µL.
5. Prepare 10x master mix; combine target cell suspension volume (for 8,000 recovered cells) with RT reagents per the 10x loading table.
6. Load Chip G: cells+master mix in row 1, gel beads in row 2, partitioning oil in row 3; run Chromium Controller to generate GEMs.
7. Transfer GEM emulsion; run RT in thermocycler (53°C 45 min, 85°C 5 min).
8. Break emulsion (recovery agent), purify cDNA with Dynabeads + SPRI.
9. Amplify cDNA (11-12 cycles); clean 0.6x SPRI; QC on Bioanalyzer (cDNA peak 400-9,000 bp).
10. Fragment cDNA, end-repair, A-tail, ligate adapters; add Dual Index PCR (sample index).
11. Double-sided SPRI size selection; final QC by Qubit + TapeStation (insert ~350-450 bp).

Independent variable: condition/genotype of source mouse. Dependent variables: number of recovered cells, median genes and UMIs per cell, mitochondrial read fraction, multiplet rate, and fraction of reads in cells. Controlled variables: target cell loading (8,000), viability threshold (>85%), cell concentration (~1,000 cells/µL), dissociation time, RT cycle parameters, cDNA amplification cycles, and gel-bead kit lot.

If a freshly dissociated mouse spleen suspension is filtered, RBC-depleted, dead-cell-removed to >85% viability, and loaded at the 10x-recommended concentration to target 8,000 cells, then GEM generation will yield single-cell transcriptomes with a low multiplet rate (~6%), high gene-detection sensitivity (>2,000 median genes/cell), and clearly separable immune populations on clustering.

Process raw BCLs with Cell Ranger (mkfastq + count) against the mouse mm10/GRCm39 reference to produce filtered feature-barcode matrices. Import into Seurat/Scanpy. QC-filter cells by nFeature (>500), nCount, and percent.mito (<10%); remove doublets with scDblFinder/Scrublet; correct ambient RNA with SoupX/CellBender. Normalize (SCTransform or log-normalize), integrate replicates (Harmony/CCA), cluster (Leiden), and annotate cell types by marker genes.

1. Low cell recovery: clogged chip or low input concentration — re-count and reload at correct concentration, ensure no clumps (re-filter through 40 µm). 2. High ambient RNA / low fraction reads-in-cells: dead cells or debris — improve dead-cell removal and process faster; apply CellBender. 3. High multiplet rate: overloaded — reduce target cells or confirm accurate count. 4. Low genes/cell: RNA degradation or over-dissociation — minimize dissociation time, keep cold, use RNase-free reagents. 5. Wetting failure / no emulsion: chip not seated properly or reagent bubbles — reload a fresh chip, avoid pipetting bubbles into wells.

Differential expression between conditions per cluster using a pseudobulk approach (aggregate counts per cell type per replicate, then DESeq2/edgeR), n=3 mice per group, negative-binomial test with Benjamini-Hochberg FDR at α = 0.05 — preferred over per-cell Wilcoxon to avoid pseudoreplication. Report log2FC and padj. Cell-proportion shifts tested with a Dirichlet/scCODA or beta-binomial model. Power depends on cell numbers per cluster; report cells/cluster and effect sizes.