The heterogeneity of cells in culture and in organisms poses a challenge for many experimental measurements. Population measurements are necessarily averages, masking the behavior of minority subpopulations and effectively blinding researchers to possibly interesting differences between cells.

Methods to study individual cells are needed in the face of cellular heterogeneity. Credit: Katie Vicari

The alternative is to make measurements on single cells. Methodologically speaking, this, too, is challenging on several fronts. Molecular analyses, whether on a particular macromolecule or at an 'omic' scale, can be difficult (or even impossible) to accomplish on the amount of material extracted from one cell. Methods with increased sensitivity are therefore in demand. Throughput is also a bottleneck. Basing firm conclusions on single-cell measurements means that one must be able to quickly and accurately analyze many cells. Finally, it is often necessary to analyze single cells in a multiplexed fashion, either because the cells exist in a heterogeneous population or because one wants to measure many parameters at the same time.

There continue to be methodological advances on all of these fronts. Mass cytometry, for instance—in which isotopes are used as antibody labels instead of fluorescent probes—considerably extends the multiplexing capabilities of flow cytometry (Science 332, 687–695; 2011). In the measurement of gene expression, digital reverse-transcriptase PCR in a microfluidics device makes it possible to simultaneously monitor the expression of hundreds of genes in hundreds of single cells. As demonstrated in a recent study of tumor heterogeneity, this can becombined with single cell sorting and with statistical clustering methods to begin to dissect the cellular subpopulations that constitute a tissue (Nat. Biotechnol. 29, 1120–1127; 2011). Microfluidics has also been at the heart of recent advances in molecular haplotyping, a measurement that is inherently made on a single cell.

As single-cell analysis is increasingly applied to ask biological questions, the demands for sensitivity and throughput will only increase, in particular for methods to read out macromolecules other than DNA and RNA. Biology is complex enough to keep single-cell methods developers in business for a while yet.