single-cell sequencing

Imagine trying to understand a city's mood by reading the average of everyone's diary blended into one. You would learn the typical sentence but lose every distinct voice — the night-shift nurse, the toddler, the protester. Bulk sequencing of a tissue does exactly this averaging: it grinds up thousands or millions of cells together and reports their pooled DNA or RNA. Single-cell sequencing instead reads each cell separately, so the distinct voices come back.

The most common form, single-cell RNA sequencing, measures which genes are switched on in one cell at a time. The trick is to physically separate cells — often by trapping each in its own tiny droplet of oil — and tag every cell's RNA with a unique molecular barcode before pooling everything for sequencing. Afterward, the barcodes let the computer sort the mixed reads back to the cell each came from. The result is a giant table: thousands of cells down one axis, thousands of genes across the other, with each entry roughly how much that gene was expressed in that cell. From this, software groups cells by their expression patterns and reveals the distinct cell types and states hidden inside a tissue.

This has transformed biology by exposing what averaging concealed: rare cell types, the trajectory a cell follows as it matures, and the surprising variability between cells that look identical under a microscope. It powers projects like the Human Cell Atlas. But the data are demanding and easy to over-read — single cells yield tiny, noisy amounts of RNA, so many genes read as zero simply because they were missed (a 'dropout'), not because they were silent. The clusters and 'cell types' it produces are statistical groupings that require careful interpretation, and the technique usually destroys the cell, giving a snapshot rather than a movie.