Single-cell eQTL mapping: how DNA variants tune genes in each cell type

What it is

Single-cell eQTL mapping finds inherited DNA variants that change how strongly a gene is expressed, measured separately in each type of cell. An eQTL, or expression quantitative trait locus, is a place in the genome where people carrying different versions of a variant tend to show different amounts of a gene's RNA; the single-cell approach asks in which cells that difference appears.

The analysis needs two kinds of data from the same people: their genotypes, and single-cell RNA sequencing, which counts RNA molecules in thousands of individual cells. Cells are sorted into types and states by their expression profiles and traced back to their donors, sometimes using each donor's own variants as a barcode when samples are pooled. A common approach sums counts per donor and cell type into a pseudobulk profile, then tests whether expression rises or falls with the number of copies of a nearby variant. Other methods model cells individually or follow expression along a continuous path, such as a cell maturing or responding to a stimulus.

How it differs from bulk-tissue eQTL studies and genome-wide association studies (GWAS)

Bulk-tissue eQTL studies measure the average RNA level across all the cells in a tissue or blood sample. That average blurs effects confined to one cell type, especially a rare one, and can be confounded when people's samples contain different mixes of cells. Single-cell mapping separates those signals, so a variant that matters only in one kind of immune cell is not diluted by all the others. A peer-reviewed study that sequenced more than five million immune cells from 1,925 people reports that inherited variation, rare as well as common, often tunes gene expression in just one or a few cell types.[1]

GWAS link variants to traits or diseases rather than to gene activity. Most variants they flag lie outside protein-coding sequence, so a GWAS hit rarely says which gene is affected, or where. Single-cell eQTLs can offer a candidate answer: if one variant appears to drive both disease risk and a gene's expression in a particular cell type, that gene and cell type become leads for follow-up. The idea extends beyond RNA. Another peer-reviewed study ties genetic variants to how open chromatin is in individual cell types and reports that this genetic signal helps an AI model map how genes are switched on.[2]

Where it is stuck

Statistical power is the main constraint. Single-cell sequencing is costly, and the unit that carries genetic variation is the person, not the cell, so even millions of cells from a limited number of donors give modest power; rare cell types and rare variants are hit hardest. Counts per cell are sparse, and decisions about how finely to divide cells into types or states can change which effects are found and make studies hard to compare.

Context is another gap. Some regulatory effects may surface only when cells are activated, stressed or at a particular developmental stage, so resting samples can miss them. As in much of human genetics, cohorts have drawn heavily on people of European ancestry, which may limit how well findings transfer. Pooling results into shared single-cell eQTL databases helps, but differing cell labels, pipelines and file formats make harmonisation difficult.

Finally, an eQTL is a statistical association, not a mechanism. Showing that a variant and a disease share a causal signal, and that the gene and cell type it implicates are the ones that matter, generally still needs further statistical work and laboratory experiments. Many disease-associated variants have yet to be matched to a convincing expression effect.