Fig. 4: Allele specific quantification of transcripts at the site of transcription suggests a transcription-based mechanism.

a An example image of data showing Xist (green) and Tsix (magenta), with brighter signal at allele specific transcription sites (surrounded by an orange box) compared to signal at individual transcripts. Colocalized signal is white. b Conceptual diagram showing the expected data. Regulation at the transcription site would result in negative correlation between Xist and Tsix levels at the site of transcription (left). No transcription-based regulation would result in no correlation at the transcription site and co-transcription would be likely (right). c The joint probability distribution of allele specific Xist and Tsix transcripts at the site of transcription (determined by total brightness / brightness of an individual transcript) in three different replica experiments. Rho and p values are determined by the Spearman correlation test, and significant negative correlation is found in all three replicas. d The joint probability distribution of the randomized data (randomized the same as Fig. 3). A two-dimensional two sided KS test⁵⁶ (p-value shown) determined significant differences between the original negatively correlated data and the randomized version in all replicas. Xist and Tsix transcripts in the site of transcription is binned by 3 transcripts each. Color code corresponds to the negative logarithm base 10 of the probability. Replica 1, n = 963 cells, 1045 Tsix transcription sites, 7437 nascent Tsix transcripts, 377 Xist transcription sites, 2936 nascent Xist transcripts; replica 2, n = 1063 cells, 867 Tsix transcription sites, 5963 nascent Tsix transcripts, 316 Xist transcription sites, 2290 nascent Xist transcripts; replica 3, n = 814 cells, 711 Tsix transcription sites, 4855 nascent Tsix transcripts, 507 Xist transcription sites, 9695 nascent Xist transcripts. Figure 4 analyzed the transcription sites for the same cells as shown in Fig. 3.