Assay

Single-Nucleus RNA-seq: sanity checks and pitfalls

Nuclei are isolated from frozen tissue and profiled with the same droplet chemistry as cells, but roughly half of the signal is unspliced pre-mRNA, so counting must include introns or you lose most of the reads. Mitochondrial transcripts should be near absent, which flips the meaning of a mito filter, and ambient RNA from lysed cytoplasm is a larger fraction of each droplet than in whole-cell data. Cell-type proportions are also distorted relative to scRNA-seq, so comparisons across the two assays need care.

Who this is for: Neuroscientists, cardiac and kidney biologists, and anyone working from frozen or hard-to-dissociate tissue. They usually inherit scRNA-seq defaults and are surprised that intronic reads, low mitochondrial fraction and ambient contamination behave differently in nuclei.