Chapter 1: Single-Cell Technologies, Droplet Microfluidics & Chemistries

Chapter 1: Single-Cell Technologies, Droplet Microfluidics & Chemistries

Single-cell RNA sequencing (scRNA-seq) has transformed biology by replacing bulk transcriptional averages with high-resolution cellular heterogeneity. Rather than measuring the mean expression of millions of blended cells, single-cell approaches isolate, barcode, and sequence individual transcriptomes, uncovering rare cell types, transitional progenitor states, and clonal dynamics.

1. Droplet-Based Microfluidics (10x Chromium & InDrop)

Modern high-throughput single-cell platforms rely on microfluidic droplet generation where aqueous cell suspensions, lysis reagents, and reverse transcription enzymes meet an immiscible fluorinated oil stream at a microfabricated junction (e.g. cross or T-junction).

In the 10x Genomics Chromium system, Gel Beads in Emulsion (GEMs) are formed. Each gel bead is coated with millions of functionalized oligonucleotides containing:

1. Illumina Read 1 sequencing primer site
2. 16 bp 10x Cell Barcode (selected from an error-correcting whitelist of ~737,000 sequences)
3. 12 bp Unique Molecular Identifier (UMI)
4. 30 bp poly-dT tract (in 3' gene expression kits) or template-switching oligo (in 5' immune profiling kits)

2. Poisson Loading Physics and Multiplet Control

To avoid encapsulating multiple cells within the same droplet (a "doublet" or "multiplet"), cells are loaded at extreme limiting dilution. The probability \(P(k)\) of encapsulating exactly \(k\) cells inside a droplet follows a discrete Poisson distribution:

P(k) = rac{\lambda^k e^{-\lambda}}{k!}

where \(\lambda\) is the average number of cells per droplet (typically \(\lambda pprox 0.05 - 0.1\)). Consequently:

• ~90–95% of droplets are completely empty (GEMs containing only bead and oil)
• ~5–9% contain a single cell ("singlet")
• <1% contain two or more cells ("multiplet")

Doublet rate scales linearly with the number of loaded cells (~0.8% per 1,000 recovered cells). If you target 10,000 recovered cells, expect a ~7–8% baseline doublet rate that must be computationally scrubbed.

3. Technology Comparison: Droplets, Plate-Based & Split-Pool

# Technology Benchmark for Single-Cell Transcriptomics:
# -------------------------------------------------------------------------------------------------------
# Technology        Throughput       Read Coverage     Cost / Cell    Best Use Case
# -------------------------------------------------------------------------------------------------------
# 10x Chromium 3'   1k - 20k cells   3' end bias       Moderate       Large-scale atlas & cell typing
# 10x Chromium 5'   1k - 20k cells   5' end bias       Moderate       VDJ immune repertoire + RNA
# SMART-seq3        96 - 384 cells   Full transcript   High           Isoform switches, RNA editing, SNVs
# Parse SPLiT-seq   10k - 100k+      Combinatorial     Low            Scale-up without microfluidic chips
# BD Rhapsody       1k - 20k cells   Microwell beads   Moderate       Targeted panels & high mRNA capture
# -------------------------------------------------------------------------------------------------------

4. Full-Length vs. Digital 3'/5' End Counting

While 10x Chromium sequences only the terminal 200–400 bp adjacent to the poly-A tail (providing digital expression counts via UMIs), SMART-seq3 generates full-length cDNA across entire gene bodies. This allows researchers to detect alternative splicing events, exon skipping, gene fusions, and somatic point mutations at single-cell resolution, albeit at higher per-cell library preparation costs.