Chapter 15: Introduction to Single-Cell RNA-Seq: 10x Genomics, Cell Ranger & Quality Control

1. From Bulk Average Expression to Single-Cell Resolution

Bulk RNA-Seq averages the transcriptional output of hundreds of thousands of heterogeneous cells, masking rare stem cell niches, drug-tolerant persister cells, and specialized tumor-infiltrating lymphocyte clones. Microfluidic droplet technologies (such as 10x Genomics Chromium) isolate individual cells inside microscopic oil droplets alongside hydrogel beads delivering two vital oligonucleotide barcodes:

2. The 10x Genomics Cell Ranger Pipeline

# Executing 10x Genomics Cell Ranger Count Pipeline
cellranger count --id=Patient_PBMC_Trial \
                 --transcriptome=/ref/refdata-gex-GRCh38-2024-A \
                 --fastqs=/data/pbmc_fastqs/ \
                 --sample=PBMC_Sample1 \
                 --localcores=16 \
                 --localmem=64

# Output Directory Structure:
# outs/
# ├── filtered_feature_bc_matrix/
# │   ├── barcodes.tsv.gz   (Identified valid cell barcodes)
# │   ├── features.tsv.gz   (Ensembl IDs and gene symbols)
# │   └── matrix.mtx.gz     (Sparse count matrix)
# ├── web_summary.html      (Interactive sequencing QC metrics)
# └── possorted_genome_bam.bam

3. Single-Cell Quality Control & Filtering in R

Before downstream analysis, three critical technical artifacts must be filtered out:

1. Empty Droplets: Ambient free-floating RNA packaged in droplets lacking intact cells (extremely low UMI count).
2. Apoptotic / Damaged Cells: Lysed cells where cytoplasmic mRNA leaked out, leaving high percentages of mitochondrial transcripts (percent.mt > 15%).
3. Doublets / Multiplets: Two or more cells captured in a single droplet (abnormally high detected gene count).

library(Seurat)
library(ggplot2)

# Load 10x Genomics filtered sparse matrix
pbmc.data <- Read10X(data.dir = "filtered_feature_bc_matrix/")
pbmc <- CreateSeuratObject(counts = pbmc.data, project = "PBMC_BioResearch", min.cells = 3, min.features = 200)

# Calculate Percentage Mitochondrial Gene Expression
pbmc[["percent.mt"]] <- PercentageFeatureSet(pbmc, pattern = "^MT-")

# Visualize Pre-Filter QC Violin Plots
VlnPlot(pbmc, features = c("nFeature_RNA", "nCount_RNA", "percent.mt"), ncol = 3)

# Apply Strict Single-Cell Filtering Criteria
pbmc <- subset(
    pbmc,
    subset = nFeature_RNA > 500 &
             nFeature_RNA < 4000 &
             percent.mt < 10 &
             nCount_RNA < 20000
)
cat("Viable single cells retained after QC:", ncol(pbmc), "\n")