Dynema method maps context-dependent eQTLs at true single-cell resolution

A preprint posted to bioRxiv on 29 August 2026 introduces Dynema (Dynamic eQTL mapping in single cells), a computational method for genome-wide identification of expression quantitative trait loci (eQTLs) at the resolution of individual cells rather than collapsed pseudobulk aggregates.

Most existing single-cell eQTL studies compress single-cell data into per-donor, per-cluster averages — a pseudobulking step that improves statistical tractability but may obscure regulatory effects that vary continuously with cell state. Dynema instead fits a Poisson model with cluster-robust variance estimators directly to single-cell count data, enabling the detection of both context-independent eQTLs and context-dependent eQTLs whose effect sizes shift along cell-state gradients.

The authors report that Dynema achieves efficient genome-wide mapping speeds and reproduces established eQTLs while additionally recovering dynamic regulatory signals missed by pseudobulk approaches. The method is framed as a tool for connecting disease-associated variants — identified through genome-wide association studies — to the specific cell states in which those variants exert their regulatory effects.

This is a preprint and has not yet undergone peer review. The statistical claims and benchmarks reported should be treated as preliminary until independent evaluation is complete.