Figure 4.
Top 20 genes, ranked by highest PSI value, have putative alterations on the protein level. (A) PSI values for alternative 3′ss were calculated by dividing the number of reads spanning the alternative 3′ss by the number of reads spanning the alternative 3′ss + the number of reads spanning the canonical splice site. (B) Hierarchical clustering of PSI values of novel 3′ss vs canonical 3′ss of the top 20 genes ranked by ascending PSI values. Only junction sites introducing novel peptide sequences were considered. (C) PSI-specific splicing patterns were observed for amino acid changes K700 and K666. (D) Fragment length analysis was used to validate RNA-based PSI measurements. (E) Validation of PSI values for 4 patients and 3 healthy controls with fragment length analysis. Experimental PSI values were calculated using relative peak heights of canonical and alternate 3′ss. No alternative 3′ splicing for UBL7 was present in patients or healthy controls. (F) Summary of differential splice junction (junction expression, columns 4 and 5) and gene expression analysis (wild-type gene expression, columns 7 and 8) of SF3B1 mutated versus wild-type patients. Differential splice junction and gene expression identified 850 and 828 significantly regulated junctions and genes, respectively (supplemental Tables 13 and 17). Normalized junction and gene read counts were extracted for the previously identified top 20 genes with highest PSI. Columns 6 and 9: log2(fold change) mutant vs wild-type for junction and gene expression. Column 10: predicted amino acid sequences introduced by the splicing defects. *P < .05, ***P < .0005 (adjusted P values). chr., chromosome; Dst., distal (distal acceptor distance − number or nucleotides between aberrant and canonical splice site); mut, mutant; wt, wild-type.