Single-cell epigenome library prep drops from a week to a day: droplet microfluidics with nanopore long reads handles about 5,000 cells at once
Synopsis
Single-cell epigenome sequencing has moved its barcoding step off the plate and into droplets. Long-read single-cell ATAC and CUT&Tag previously ran well by well, needing at least a week of library work for about 5,000 cells at high cost. Now nuclei are co-encapsulated with barcoded beads on a commercial microfluidic chip and read on Nanopore: about 5,000 cells per run, one day of library prep instead of a week, and roughly $0.39 per cell instead of $1.98. In GM12878, 10,000 input nuclei produced 66 Gb, with 82.8% of reads on valid cell barcodes.
Interpretation
Barcoding for long-read single-cell epigenomes now happens inside droplets rather than in separate wells, lifting one experiment to roughly 5,000 cells. Plate-based methods such as scNanoATAC-seq and scNanoSeq-CUT&Tag processed cells well by well, so a library for about 5,000 cells took at least a week and cost enough to keep studies small. For about 5,000 cells the droplet workflow cut library construction to one day and lowered total cost per cell including sequencing from about $1.98 to about $0.39; in GM12878, 10,000 input nuclei produced 66 Gb, 82.8% of reads carried valid cell barcodes, and 8,057 cells passed a filter of more than 500 unique fragments.
The enzymatic tagging of open chromatin or histone marks stays the same; only the barcode is now delivered in droplets. dscNanoATAC transposes nuclei with Tn5 carrying two identical adapters that pair with oligonucleotides on microfluidic beads, so after encapsulation the two fragment ends prime off each other and acquire the same cell barcode and unique molecular identifier; dscNanoCUT instead binds a histone mark with an antibody and lets protein G-fused Tn5 do the tagmentation. Hairpin priming accounted for 82.5% of dual-barcode reads against 17.5% from a second primer, longer extension pushed amplicons to about 4 kb for Nanopore sequencing, and a mixture of four human cell lines resolved into cleanly separated clusters with a cluster-separation score (silhouette coefficient) of 0.90 across 7,064 cells.
Perspective
The demonstrated scope is cultured human cell lines (GM12878, K562, H9, HEK293T), mouse NIH3T3, and mouse cerebral cortex and hippocampal formation tissue, and it was validated on only one commercial microfluidic platform, MobiCube; redesigning the Tn5 adapter and amplification primer to match bead oligonucleotides would extend the same strategy to other droplet platforms.
The methods currently capture epigenomic data only and lack matched gene expression, so they cannot directly assess how accessibility and histone modification affect transcription; at the same sequencing depth long reads yield fewer fragments than short-read methods. AS-C.Origami allele-specific predictions are treated as reliable only when at least 80% of the 10 kb bins in a 2 Mb window contain one or more SNPs, and such regions cover 47% of the GM12878 genome (2,820 of 6,004 windows).