Fig. 4

Landscape of alternative promoter usage in the Iso-seq data. a Correlation matrix between the detected isoforms from Iso-seq data with the promoter activity estimated from Illumina short-read RNA-seq for isoforms with length less than 2 kb. The highest correlation coefficients are always observed between the predicted promoter activity from the Iso-seq and Illumina data from the same cell line. b Number of identified Iso-seq genes and promoters. Twenty-five percent of all genes have multiple promoters. c Example of known and novel promoters. Promoters are considered known if at least one FSM transcript is initiated from the promoter site, and novel if no FSM transcript is initiated from the promoter site. Gene promoter with the higher average activity is additionally assigned as major promoter and all other promoters for the same gene are assigned as minor promoters. d Assignment of major and minor promoter from short-read RNA-seq data. Novel promoters are often minor promoters. e Schematic representation of the relationship between 5′ UTR, CDs, and 3′ UTR regions. On average, usage of alternative promoters modifies about 22–24% of coding regions. f Example of how alternative promoter usage can modify CDs (PSMB4) and 5′ UTR (MRPL28)