Fig. 1. Severity of splicing dysregulation correlates with PCa evolution.

a Schematic illustrating the spectrum of PCa development, therapy resistance and metastatic progression, and related datasets (Oncomine and RNA-seq with reference PMID provided) used for pairwise comparisons. The spectrum of PCa progression is indicated by Normal → Pri-PCa → CRPC-Ad → CRPC-NE, and the relationship of each dataset to specific PCa progression stages is indicated by various arrows. For example, we have previously shown that normal prostate luminal and basal cell profiles molecularly resemble pri-PCa (blue arrow) and aggressive subtypes (i.e., CRPC-Ad and CRPC-NE, red arrow), respectively⁵. On the other hand, prostate tumors before and after ADT are related to pri-PCa and CRPC-Ad (dark yellow), respectively⁵. b Five main types of AS patterns analyzed in the present study. A3, alternative 3′ splice sites; A5, alternative 5′ splice sites; MX, mutually exclusive exons; SE, exon skipping; IR, intron retention. c–l Alterations in AS landscape during PCa development and progression. Two related datasets are interrogated and compared for each PCa stage. Shown are splicing patterns and the number of DSEs decoded by rMATS. DSEs, differentially spliced events. See Supplementary Data 3 for details.