Fig. 2: Performance of scRCAT-seq2.

a Coverage of scRCAT-seq2 data on transcripts of single ES cells (n = 6, A total of 6 replicates of scRCAT-seq2 were performed, each time using one ES cell with SIRV-Set 3). The displayed values represent the averaged coverage of reads with shaded areas indicating standard deviations. b In SIRV-set 3, the gene SIRV1 is demonstrated as an example, where the full-length transcriptome was identified and constructed through scRCAT-seq2. The center line: median; boxes: first and third quartiles; whiskers: 5th and 95th percentiles. c Sensitivity of scRCAT-seq2 in identifying transcripts in SIRV-Set 3, with 88% (61 isoforms were characterized from 69 isoforms) of the transcripts detected. Error bars represent the standard deviation of the mean (n = 6, A total of 6 replicates of scRCAT-seq2 were performed, each time using one ES cell with SIRV-Set 3). d Correlation scatter plot between the expected (x-axis) and observed abundance (y-axis) of ERCC transcripts (log-transformed, Pearson correlation coefficient r = 0.97, n = 92, A total of 92 ERCC transcripts). e Correlation scatter plot of ERCC transcripts between replicates (Pearson correlation coefficient r = 0.995, n = 6). f Correlation scatter plot of gene expression levels in hESC samples between replicates (Pearson correlation coefficient r = 0.936, n = 6). g The number of isoforms detected using scRCAT-seq2 (hESC, n = 6; HEK293T, n = 6), SCAN-seq (HEK293T, n = 6; K562, n = 6), or scISOr-seq (hESC, n = 2), versus cost. Displayed is the mean number of transcripts with 95% confidence intervals shaded. h Comparison of the number of transcripts detected by scRCAT-seq2 (n = 36, A total of 36 ES cells and 293T cells with the same sequencing cost), SCAN-seq (n = 240, A total of 240 K562 cells and 293T cells with the same sequencing cost), scISOr-seq (n = 2, A total of 2 ES cells with the same sequencing cost), and scRCAT-seq (n = 2, A total of 2 ES cells with the same sequencing cost) at the same cost in low-throughput sequencing. Significance was computed using two-sided t-test. The center line: median; boxes: first and third quartiles; whiskers: 5th and 95th percentiles. i Comparison of the number of isoforms detected by scRCAT-seq2, nanopore, and scRCAT-seq in high-throughput sequencing of retinal organoids; comparison of the number of isoforms detected by scRCAT-seq2 and HIT-scISO-seq in high-throughput sequencing of monkey corneal limbal epithelium cells (Limbus). Significance was computed using two-sided t-test. The center line: median; boxes: first and third quartiles; whiskers: 5th and 95th percentiles. Source data are provided as a Source Data file.