Fig. 2: 5′-C transcripts are most enriched in poorly differentiated and proliferative cancer types.
Dual initiators (promoters with >1 TPM for both YC and YR transcription in the majority of datasets (>30 out of 59)) were identified across all selected FANTOM5 cancer and healthy tissue CAGE datasets (n = 3,475). The relative expression of the YC versus YR component of transcription for each dual initiator was calculated and compared between matched cancer and healthy tissues. a, Table of cancer samples ordered by mean log2 fold change (log2(FC)) in the ratio of YC to YR transcription of dual-initiator promoters between cancer and matched healthy tissue. P value (paired two-tailed t-test) of this expression change is also shown (ns, not significant; *P < 0.05; **P < 0.01; ***P < 0.001; full list of P values available in the Source data). This table is color-coded to show cancers where YC transcription at dual initiators is significantly enriched (orange), unchanged (gray) or depleted (green) relative to the matched healthy tissues. b, The differentiation status of each cancer sample was identified where publicly available. They were then separated into undifferentiated, moderately differentiated and well-differentiated cancer types, and the distribution of each sample’s mean log2(FC) in YC:YR transcription (as calculated in a) was plotted for each group. The distribution of TP53-mutant samples is also shown by the color of each plot point. c,d, Biological process ontology of genes specifically upregulated in YC-enriched cancers (n = 132) (c) and YC-depleted cancers (n = 144) (d). FDR, false discovery rate.
