Fig. 2: Homotrimer UMIs enhance differential expression accuracy in single-cell sequencing.

a, Human Jurkat and mouse 5TGM1 cells were mixed for encapsulation and cDNA synthesis using 10X chromium followed by nanopore sequencing. Each dot represents an individual cell. b, A UMAP of 10X chromium data showing the integration, clustering and annotation of human and mouse cells following 20 and 25 cycles of PCR. c, A density plot for the 10X chromium data showing the log₁₀ density of the number of UMIs following 20 and 25 cycles of PCR. The dotted line shows the maximum density for each condition. d, A UMAP showing the expression of ENSMUST00000034966 between libraries amplified following 20 and 25 PCR cycles. e, A UMAP showing the expression of ENST00000532223 between libraries amplified following 20 and 25 cycles of PCR. f, A schematic showing the homotrimer UMI drop-seq library preparation approach and template switching attachment of a homotrimer CMI to single-cell captured mRNAs. g, Drop-seq libraries were sequenced using the Flongle sequencing device, graphs show the percentage of reads that have an accurate CMI following amplification of the same library using 20, 25, 30 and 35 cycles of PCR before and after homotrimer correction. Error bars are the s.d. of three independent experiments. h, Barnyard plots showing the expression of mouse and human cells following 20 and 25 cycles of PCR and sequencing using a PromethION sequencing device. Each dot represents an individual cell. i,j, UMAP plots showing the transcript expression of ENST00000330494 following monomer-based UMI-tools demultiplexing (i) and homotrimer-based demultiplexing (j). In each UMAP plot, each dot represents a cell. Data shown in g were collected in a single sequencing run and n = 1.