Table 1.

Key scRNA-seq and snRNA-seq method adaptations

Method	Cells	Concept	Advantages	Disadvantages	Reference
STRT-seq (single-cell tagged reverse-transcription sequencing)	1 × 102	Indexed template-switching oligos used for RT-based barcoding of cells in individual wells	First barcoding strategy allowing for multiplexing of multiple cells Multiplexing minimizes well-to-well technical biases	5ʹ bias Manual single cell isolation	[4]
CEL-seq (Cell Expression by Linear amplification and sequencing)	5 × 102	Barcoded RT primers allows early pooling before IVT amplification	Linear amplification preserves relative abundances of mRNA transcripts Detects more genes than STRT-seq Higher sensitivity than STRT-seq	3ʹ bias High-abundance transcripts bias Manual single cell isolation	[7]
SMART-seq (Switching Mechanism at 5ʹ End of RNA Template)	1 × 102	Full-length cDNA amplified following a template-switching reverse transcription reaction	Full-length cDNA Allows splice isoforms to be discerned	Lack of strand specificity Selective for polyadenylated RNA Increased labour compared to methods with early multiplexing	[13]
SMART-seq2 (Switching Mechanism at 5ʹ End of RNA Template)	102–103	As above	As above Increased cDNA yield from single cells, higher sensitivity, fewer technical biases, less variability, and less expensive compared to SMART-seq Initial steps are used by many scRNA-seq workflows	As above Transcript length bias (inefficient with mRNAs > 4 Kb) High-abundance transcript bias	[14]
SMART-seq3	102–103	As above but with UMI incorporated into template-switching oligo for in silico isoform analysis.	Optimized steps leads to improved cDNA yields and library complexity Strand specific Improved ability to discern isoforms and allelic expression	As above Transcript length limitation to in silico isoform analysis	[15]
Quartz-seq/Quartz-seq2	102–104	Poly-A tailing and 2nd strand synthesis instead of template-switching oligo for full-length transcript amplification	Highly reproducible cell transcriptomes Increased library complexity leads to reduced sequencing requirements High UMI conversion rate for quantitative studies.	High mRNA GC content reduces detection 3ʹ bias Bias towards shorter transcripts	[18,132]
Fluidigm C1	102–103	Individual cell capture and library processing in commercially available integrated microfluidics circuits	Automated processing minimizes technical bias Full length sequencing Circuits optimized for different cell sizes	Specialist equipment required	[6]
MARS-Seq (massively parallel single-cell RNA-sequencing)	> 103 cells	Automated, multi-step barcoding of cells FACS sorting of cells into well, three levels of barcoding	Automatization of single cell isolation with FACS Increased throughput possible High degree of multiplexing	Requires specialist equipment (FACS, liquid handler) Lack of strand specificity 3ʹ bias	[12]
Droplet based scRNA-seq (i.e. Drop-seq/inDrops)	103–106	Microfluidics capture of cells in oil droplets, RT based barcoding i) on mRNA capture beads (Drop-seq) or ii) in droplets (inDrops)	Parallel processing of large number of cells for increased scalability at reduced cost	3ʹ bias Requires custom microfluidics system Low mRNA capture efficiency with inDrops	[22,43]
SPLiT-seq (Split-pool ligation-based transcriptome sequencing)	> 105	In situ barcoding of RNA via multiple sequential split-pool reactions	Compatible with fixed cells or nuclei Efficient sample multiplexing No customized equipment required Avoids non-trivial single cell isolation/sorting	Labour intensive Low complexity per cell in current studies 3ʹ bias	[24]
sci-RNA-seq (single-cell combinatorial indexing RNA sequencing)	5 × 104	In situ mRNA barcoding followed by second barcoding step during PCR amplification	Compatible with fixed cells or nuclei Efficient sample multiplexing Scalable via extra tagmentation barcoding, or use of 384-well plates FACS step helps eliminate doublets	Labour intensive 3ʹ bias	[46]
Scifi-seq (single cell combinatorial fluidic indexing)	103–106	In situ pre-indexing of cell transcriptomes followed by scRNA-seq with droplet based sequencing	Multiple cells per droplet increases cell/nuclei throughput ~15-fold. Easy multiplexing Faster than multi-round combinatorial indexing Efficient reagent use and sequencing costs High complexity per cell	Non-trivial optimization 3ʹ bias Requires custom microfluidics system	[47]
Pico-well based sequencing	>104	PDMS-based printing of >10,000 pico-wells to facilitate the gravitational capture of individual cells and mRNA capture beads in high throughput	Cost-effective Readily scalable Recent versions improve cell separation and reduce cross-contamination	Requires custom fabricated chips 3ʹ bias	[40]
snRNA-seq (i.e. SNS, sNuc-seq)		Single nuclei used instead of whole cell SNS uses fluidigm C1 system sNuc-seq uses Smart-seq2	Study of difficult to dissociate cell types (e.g. neural tissue, archived tissue) Full length sequencing	Low mRNA capture due to rapid nuclear export following poly-adenylation Missing information from cytoplasmic transcriptome FACS dependent	[58,133]
Droplet based snRNA-seq (i.e. DroNc-seq, snDrop-seq)	5 × 104	Massively parallel single nuclei profiling with droplet technology	Study of difficult to dissociate cell types (e.g. neural tissue, archived tissue) Parallel processing of large number of nuclei for increased scalability at reduced cost	Low mRNA capture due to rapid nuclear export following poly-adenylation Missing information from cytoplasmatic transcriptome 3ʹ bias	[23,57]