There are many reasons for analyzing the gene expression profiles of individual cells rather than a pooled population. But scaling up single-cell transcriptomics to analyze multiple cells in parallel has been challenging. “One can only analyze a few, or maybe 100 cells at a time,” says Donald Zack of Johns Hopkins Hospital in Baltimore.
“There’s a need for techniques that are unlimited in terms of the number of cells,” says Marc Kirschner of Harvard University. In some instances you need increased statistical power, he says. “But also, it’s often the case that you’re looking for a needle in a haystack”—that is, extremely rare cells.
Kirschner’s laboratory and that of Steve McCarroll, also at Harvard, have now independently broken the scale barrier for single-cell analyses. And surprisingly, the teams achieved their feats in similar ways despite only learning of each other’s approach well along in development. Both groups used microfluidic devices to create thousands of nanoliter-size aqueous droplets housing one cell and one set of barcoded primers each. The primers’ barcode sequences were unique to each droplet such that after reverse transcription and sequencing of the RNAs the researchers could identify the cells from which each transcriptome originated.
“The innovation was using an aqueous-oil emulsion,” says Zack, describing the oil-suspended droplets that essentially form hundreds of thousands, even millions, of tiny temporary test tubes. “Instead of needing a well or a little spot for each cell, now they could, by many orders of magnitude, increase throughput,” he says. Such droplets have been used for single-cell PCR, but the new techniques are the first to apply the technology to transcriptomics.
Kirschner’s team has used its technique, called inDrop, to study variations in seemingly uniform populations of mouse embryonic stem cells, while McCarroll’s team, using the Drop-seq approach it designed, has identified known and novel cell types in the mouse retina. As both researchers assert, however, the number of possible applications is vast. (Cell, 161:1187-1201, 2015 and Cell, 161:1202-14, 2015)
|TECHNIQUE||BARCODED PRIMERS||DROPLET PRODUCTIOJN||AMOUNT OF CELLS SEQUENCED|
|Drop-Seq||16.8 million possible unique primers synthesized on surface of resin beads||Millions per hour. About 1 percent contain both a bead and a cell.||Many droplets contain a cell but no bead, so only a subset of cells is sequenced.|
|inDrop RNA sequencing||147,456 possible unique primers covalently linked to hydrogel microspheres||Approximately 100,000 per hour. Nearly 100 percent contain microspheres and 10 percent also contain a cell.||Roughly 90 percent of cells are captured and sequenced. Good approach for samples with limited cell numbers.|