Extended Data Fig. 1. Performance and quality control analyses of scNT-Seq.
a. Scatterplots showing the number of detected gene per cell (y-axis, upper panel) or UMI per cell (y-axis, lower panel) as a function of aligned reads per cell (x-axis) between 4sU (red, 462 cells), TFEA (blue, 211 cells), and 4sU/TFEA (green, 578 cells) experiments. 4sU, cells labeled with 4sU (100 μM, 4 hours (h)). TFEA, beads treated with TFEA/NaIO4 chemical reaction. 4sU/TFEA, cells labeled with 4sU and beads treated with TFEA/NaIO4 chemical reaction. The fitted lines of different experiments were shown. The predicted numbers of gene or UMI detected per cell at matching sequencing depth (50,000 aligned read per cell) are shown on the top.
b. Shown are all transcripts (with unique UMIs) for the ACTG1 gene from one untreated control K562 cell (upper panel) and one TFEA/NaIO4-treated cell (lower panel). Grey circles denote uridine sites without T-to-C substitution, and “X”s denote sites with T-to-C substitutions. The read coverage for each T-to-C substitution is color-scaled. All 9 sequencing reads of the 2nd UMI (in red box) from the TFEA/NaIO4-treated cell are highlighted below.
c. Bar plot showing nucleotide substitution rates in mESCs with different labeling parameters (100 μM 4sU for 4 h or 200 μM 4sU for 24 h) and sample processing methods (freshly isolated versus methanol fixation followed by cyro-preservation and rehydration with two different rehydration buffers: PBS-based versus SSC-based). A sample (100 μM 4sU for 4 h) that was not treated with TFEA/NaIO4 served as the control.
d. Scatterplots showing Pearson’s correlation between two biologically independent replicates of mESCs (rep1: 427 cells and rep2: 733 cells). The expression levels of new (n = 10,925 genes) and old (n = 14,496 genes) transcripts were quantified as natural log transformation of (TP10K + 1).