Extended Data Fig. 2: Barcode detection efficiency.
From: Visualizing DNA folding and RNA in embryos at single-cell resolution
a, The mean detection efficiency of each barcode for all BX-C probesets. Error bars represent s.e.m. n = 8,801 cells, n = 94,300 cells, and n = 15,230 cells for the 10-kb, 3-kb, and 2-kb probesets, respectively. b, Box-and-whisker plots comparing the distribution of detection efficiencies across barcodes between two different embryos. N.S. indicates not statistically significant (two-sided Wilcoxon test, P values indicated). For the boxplots in b–e, the red line marks the median, the upper and lower limits of the boxes mark the interquartile range, and the whiskers extend to the farthest data points within 1.5 times the interquartile range. Notches denote the 95% confidence interval around the median. Sample size (n) is indicated. c, Comparison of the distribution of detection efficiencies between cells of different embryos. d, Comparison of the detection efficiency across barcodes between anterior and posterior cells. e, Comparisons of the detection efficiency across all cells between anterior and posterior cells. Results are shown for the 10-kb, 3-kb and 2-kb probesets as indicated. f, Correlation between normalized distance data for each body segment as shown in Fig. 4 and a modified dataset in which 50% of detected barcodes were removed at random to simulate a high missed detection rate. The correlation coefficients are indicated (Pearson), computed using all unique combinations of 52 barcodes. g, Bar graph comparing the mean correlation between matched segments in the original and downsampled data shown in f compared to the mean inter-segment correlation. The correlation coefficients are indicated (Pearson), computed using all unique combinations of 52 barcodes. All 15 matched segment pairs shown in f are plotted compared to all 210 inter-segment pairs. This illustrates that the difference between the substantially downsampled data and the original measurements are much smaller than the typical differences between segments, demonstrating the robustness of the segment-specific findings to missing data. Error bars denote s.e.m. h, Histogram showing the detection efficiency per cell for each probeset.
