Skip to main content
Premium Trial:

Request an Annual Quote

ABRF Single-Cell RNA-Seq Platform Comparison Finds Similar Performance, Drawbacks for Each

MYRTLE BEACH, SC (GenomeWeb) –Different single-cell RNA sequencing platforms can detect genes equally well, but none is without error, according to a recent comparison study by researchers from the Association of Biomolecular Resource Facilities.

ABRF's Genomics Research Group (GRG) tested a range of approaches for single-cell RNA sequencing, with the aim of providing researchers with information and guidance on the various platforms. The research group compared how platforms from Fluidigm, WaferGen (which is part of Takara Bio), 10x Genomics, and Illumina/Bio-Rad performed on a set of treated and untreated breast cancer cell lines and examined their differences in terms of ease of use, cost, and time.

The goal was to help researchers decide what platform is best for a particular project, according to John Ashton of the University of Rochester, who presented the results of the comparison here at the ABRF annual meeting this week.

The research group analyzed SUM149PT cells that were treated either with the histone deacetylase inhibitor TSAor with DMSO on the WaferGen iCell8, Fluidigm C1, 10x Chromium Controller, or Illumina/Bio-Rad ddSEQ platforms. The group tested two versions of Fluidigm's C1, the 96 IFC and the HT-IFC. There were about 4,000 cells per arm, except for Fluidigm, Ashton noted, as the Fluidigm approaches have lower cell capture capacities, 96 cells or 400 cells, respectively.

At the same time, the researchers performed bulk RNA sequencing of the cells to serve as their "ground truth." For gene detection, the platforms performed roughly the same, Ashton said, though not as well as bulk sequencing. He added that the Fluidigm platform performed slightly better, though the others were within a similar range. None of the platforms reached gene detection saturation at the sequencing depths tested, he noted, suggesting they had similar sensitivity under the testing conditions.

However, Ashton said that "none of the technologies were without some error." For example, he and his colleagues uncovered some detection biases, notably due to GC content or gene length.

The single-cell approaches varied in terms of capture efficiency, according to a poster presented at the meeting. The 10x platform, for instance, had a 65 percent capture efficiency in the study, while the efficiency was only 30 percent for the WaferGen platform and just 3 percent for the Illumina/BioRad platform.

Overall, the various single-cell RNA sequencing approaches could all distinguish between cells that had been treated and those that had not, he added. However, some platforms had more outliers than others and some were better able to resolve the treatment groups than others.

Still, what platform is used does matter, he noted, as the research group uncovered low concordance for differentially expressed genes. "You don't want to jump between platforms," he said.

The GRG also compared its results to the Broad Institute's Cancer Cell Line Encyclopedia, which has characterized a number of human cancer cell lines, to gauge how well the single-cell approaches could detect high and low abundance genes. To his surprise, Ashton said, the platforms were very good at detecting lowly expressed genes.

The different single-cell RNA sequencing platforms also varied in their ease of use and turnaround times. According to the GRG poster, the 10x and Illumina/BioRad approaches were considered to be easy to use by testers, while the Fluidigm HT approach was thought to be difficult to use. Likewise, the Fluidigm HT-IFC approach also took the longest, 12 hours, while the Fluidigm 96 IFC approach took 10 hours, as did the Illumina/BioRad and 10x approaches. The WaferGen method took only seven hours.

Cost also varied. At the high end, the WaferGen approach cost $3,100 to go from cell capture to library, which came out to about $3 per cell, while the Illumina/BioRad approach cost $1,200 from cell capture to library, which worked out to between $1 and $4 per cell. The Fluidigm HT-IFC approach, meanwhile, cost $2,500 to go from cell capture to library, or between $3.13 and $6.25 per cell and its 96 IFC approach cost $2,000, or $20.83 per cell. The 10x approach cost $1,500 from cell capture to library, or $.15 to $1.50 per sample.

Ashton noted that the study is subject to a number of limitations, including batch effects, a source of technical variation that occurs when samples aren't run all at once.

The findings led the research group to conclude that each single-cell RNA sequence approach has its tradeoffs, which Ashton said researchers need to consider when choosing which to use for their projects. The results of the comparison are to be published soon.

The Scan

Sensor to Spot SARS-CoV-2 Exposure

Yale researchers have developed a sensor that gauges SARS-CoV-2 exposure, according to the New Haven Register.

Watching BA.2

A subtype of the Omicron SARS-CoV-2 variant has been found in a number of countries, the Washington Post says.

Vermont Weighs Genetic Discrimination Bill

VT Digger reports Vermont legislators are considering a bill to bar insurer from denying customers coverage based on genetic information.

PNAS Papers on Convergent Pathogen Effectors, Lineage-Related RNA Mutations, Desiccation-Tolerant Grass

In PNAS this week: convergent evolution of pathogen effectors, algorithm to identify cell lineage features, and more.