Factors to Consider When Selecting an scRNA-seq Method

The relatively recent advent of single-cell RNA sequencing (scRNA-seq) has allowed questions until recently unanswerable to be probed by the non-expert, to depths heretofore nearly unimaginable. With scRNA-seq moving from the expert lab into the wider community, researchers are now faced with the enviable task of deciding which of the multiple methods and platforms now available will best meet their needs.

This article will look at the factors to consider when selecting an scRNA-seq method, including the research question, the sample type, the number of cells to be analyzed, and the desired level of resolution.

How and why single cell?

Bulk RNA sequencing came to prevalence in the mid-2000s. Yet because it measures the average RNA expression across the tissue, it masks any cellular heterogeneity. Bulk RNA-seq is very good at distinguishing a tumor from healthy tissue, say, but could not individuate the type of T cells infiltrating that tumor from each other.

As time went on, certain other technological developments came along that helped to establish sequencing at a single-cell resolution, explains Arpita Kulkarni, Associate Director of the Single Cell Core at Harvard Medical School. These included the ability to amplify the RNA content of a single cell to provide enough material to sequence; barcoding, so that all the material that came from a single cell (be it RNA, DNA, or something else) gets a unique cellular ID for tracking, and can therefore be pooled or multiplexed with other samples with unique IDs and processed together; in vitro transcription, allowing for linear (as opposed to exponential and thus biased) amplification; and cheaper, higher throughput sequencing.

scRNA-seq generally starts with the isolation and capture (partitioning) of single cells. This is followed by steps such as reverse transcription, amplification, barcoding, and library preparation. The library is then sequenced and the data analyzed to, among other things, reveal which transcripts are co-expressed in the same cell and at what level.

scRNA-seq allows researchers to study how gene expression changes from one cell type to the other. It allows for tracing the developmental pathways in oncology and immunology, for example. Another important application is to create a high-resolution catalog of cells, known as an atlas.

Move them through

The throughput and scale of experiments are important factors in which platform to use. “This can entail both the number of cells to be analyzed per biological sample, but also the number of samples to be processed at a given time,” says Angela Churchill, Product Marketing Manager, Single Cell Applications at 10x Genomics. “Factors like sample heterogeneity, the scope of a specific research question, and/or the rarity of the cell population of interest can all influence the desired throughput for an experiment.”

Some research questions do not require massive, large-scale experiments. Low-throughput methods often use microplates to partition a few hundred to a few thousand cells into individual wells—by FACS sorting, laser microdissection, or even manually, using pipettes. “These require pretty simple equipment” and reagents found in most molecular biology labs, Kulkarni notes. The protocols are “very cost effective and not super labor-intensive.”

Other questions need more data to answer them. The microfluidics-based 10x Chromium is the market leader in high-throughput platforms “for a whole bunch of reasons,” says Kulkarni. Among these she cites “really strong customer support,” “well written-out protocols that are quite simple to follow,” and “everything is streamlined.”

Technologies have arisen “that have tried to circumvent the use of microfluidics, but are still high throughput,” using common lab equipment, she points out. Among these are Parse Biosciences’ Evercode platform, which relies on sequential rounds of combinatorial barcoding to uniquely label the cDNA from each cell. And with Fluent Biosciences’ PIPseq, cells are combined with particle-templated instant partitions (PIPs) to begin the barcoding and library preparation.

The number of samples being profiled is increasing, as is the number of cells per sample. In fact, the average size of data sets found in publications has been increasing exponentially, points out Charlie Roco, Co-founder and CTO at Parse. It is important to have a product that can scale with the size of the experiments.

Data quality

Along with scalability, Roco cites data quality as the most important consideration when choosing a platform. “Which one should I use to continue to my larger projects? How many unique genes and transcripts can I detect at a set sequencing depth? How many cells do I need to throw out because they’re essentially contaminated, or two cells (a doublet) instead of one?”

He also noted that “ambient RNA”—from cells that have broken open in the single-cell suspension, and their RNA encapsulated along with the captured cell—can be falsely attributed to the captured cell. Parse’s protocol allows them to “wash the cells, because you still have whole [fixed and permeabilized] cells intact after the labeling.”

Flexibility

Not every sample type is compatible with the standard scRNA-seq assay. Cells such as megakaryocytes are too large for most microfluidic platforms, for example, while the salinity and osmolarity requirements of some marine organisms may interfere with some of the downstream molecular biology steps. Neuronal tissue “is not the best tissue to do live experiments on,” and the high fat content of adipose tissue can inhibit efficient reverse transcription to cDNA, says Kulkarni. These types of tissues are best handled by flash freezing the tissue, extracting the nuclei, and using the nuclei in place of the live cell. “The downstream protocol for library preparation remains the same.”

Platforms such as Parse do not use microfluidics, and work with fixed cells, so can handle a broad range of sample types. On the other hand, platforms such as Chromium offers a broad range of assays that are “compatible with a variety of different sample inputs including PFA [paraformaldehyde] and FFPE [formalin-fixed, paraffin-embedded] fixed tissues, fresh/frozen samples, and suspension cells/nuclei,” notes Churchill.

One advantage of using non-live samples is that it gives researchers the ability to collect and store samples to be run as batches whenever convenient, allowing, for example, collaborators to use a central lab for processing.

The future is now

Not only are companies innovating products that allow for different inputs, but different outputs, such as chromatin accessibility, T cell receptor genes, or lipid profile, can be interrogated at the single-cell level—sometimes at the same time as another omic.

“The next wave is doing spatially resolved single-cell sequencing,” Kulkarni notes. “People are interested in doing in situ single-cell sequencing, such that you don’t disturb the three-dimensional structure of the tissue. So, a whole bunch of companies with technologies are now coming up that do spatially resolved single-cell transcriptomics, with the aim to integrate data from other, parallel, single cell or epigenomics, or all of these different multi omic studies.”

Key Takeaways

• Selecting Optimal scRNA-seq Methods: As scRNA-seq gains wider traction, researchers face the challenge of choosing the most suitable method for their needs.
• Unveiling Cellular Heterogeneity: scRNA-seq's single-cell resolution enables exploration of previously inaccessible questions about cellular diversity.
• Key Considerations for Method Choice: Researchers must weigh factors like research objectives, sample types, throughput demands, data quality, and platform flexibility when selecting scRNA-seq methods.
• Diverse Sample Compatibility: Different sample types, from fixed cells to complex tissues, can be analyzed using various platforms, offering broader applicability and versatility.
• Next-Level Capabilities: Beyond gene expression, evolving scRNA-seq technologies allow for probing chromatin accessibility, T cell receptor genes, lipid profiles, and even spatially resolved transcriptomics.
• Future of Single-Cell Analysis: Ongoing innovations promise spatially resolved single-cell sequencing, integrating data from multiple omics studies and revealing three-dimensional tissue structures.