FD-seq – fixed droplet RNA sequencing of paraformaldehyde-fixed single cells

Single-cell transcriptomic studies that require intracellular protein staining, rare cell sorting, or inactivation of infectious pathogens are severely limited. This is because current high-throughput single-cell RNA sequencing methods are either incompatible with or necessitate laborious sample preprocessing for paraformaldehyde treatment, a common tissue and cell fixation and preservation technique.

Researchers from the University of Chicago have developed FD-seq (Fixed Droplet RNA sequencing), a high-throughput method for droplet-based RNA sequencing of paraformaldehyde-fixed, permeabilized and sorted single cells. The researchers show that FD-seq preserves the RNA integrity and relative gene expression levels after fixation and permeabilization. Furthermore, FD-seq can detect a higher number of genes and transcripts than methanol fixation. They first applied FD-seq to analyze a rare subpopulation of cells supporting lytic reactivation of the human tumor virus KSHV, and identify TMEM119 as a potential host factor that mediates viral reactivation. Second, the researchers found that infection with the human betacoronavirus OC43 lead to upregulation of pro-inflammatory pathways in cells that are exposed to the virus but fail to express high levels of viral genes. FD-seq thus enables integrating phenotypic with transcriptomic information in rare cell subpopulations, and preserving and inactivating pathogenic samples.

Benchmarking and validation of FD-seq

a Bar plots showing total RNA yield from bulk live cells, and bulk fixed cells that underwent cross-link heat reversal (1 h at 56 °C) with or without 40 U/mL of proteinase K. Data are presented as mean ± standard deviation. n = 3 technical replicates. b Species-mixing plots showing the single-cell capture efficiency of Drop-seq and FD-seq. The multiplet rate for live cells and fixed cells were ~0.5% and ~1%, respectively. Human BC3 cells were combined with mouse 3T3 cells at equal concentration, and processed with Drop-seq or FD-seq. See the “Methods” section for more details. c Violin plots and box plots showing the number of detected genes in live and fixed cells for each species. For this analysis, only cells with at least 1500 transcripts were considered, and 1000 transcripts were randomly sampled from each single cell. The white dots inside the violin plots represent the median of the data, the black boxes represent the first and third quartiles, and the black lines represent the values 1.5× the interquartile range beyond the first and third quartiles. n = 157 and 164 single cells for live and fixed human samples. n = 150 and 267 single cells for live and fixed mouse samples. d Comparison of the normalized expression level of each gene between live and fixed cells for each species (see “Methods” section). Each dot represents the average expression level of a gene, and the red line indicates the line y = x. The plots also show the Pearson’s correlation coefficient ρ of the log-normalized gene expression level between live and fixed cells for each species.