Figure 3: Validation of differential alternative splicing of oncogenes and tumour suppressor genes in AA versus EA PCa specimens.

(a) Schematic representation of DS events within the indicated genes in AA PCa (top dashed lines connecting exons) and EA PCa (bottom dashed lines) based on alternative splicing ANOVA model. Closed arrowheads below exons represent primer location in qRT-PCR validation of alternatively spliced transcripts. (b) Representative RT-PCR results validating race-specific/-enriched variant transcripts in either AA PCa or EA PCa specimens. Shown are the RT-PCR results for the AA-specific/-enriched variants PI3KCD-S, FGFR3-S, TSC2-S, ITGA4-L, MET-L, NF1-L, BAK1-L and RASGRP2-b; and EA-specific/-enriched variants PIK3CD-L, FGFR3-L, TSC2-L, ITGA4-S, MET-S, NF1-S, BAK1-S and RASGRP2-a. Each lane represents an RT-PCR result from an independent PCa specimen that was also interrogated in exon array experiments. RT-PCR of EIF1AX and PPA1 transcripts served as loading controls. The qRT-PCR results are summarized in Supplementary Fig. 4. Unprocessed RT-PCR images are shown in Supplementary Fig. 11. (c) Additional validation and quantification of the ratio of the AA-enriched PI3KCD-S (short) variant and the race-independent PI3KCD-L (long) variant in a separate cohort of PCa patient specimens. RNA was isolated from n=32 AA PCa and n=30 EA PCa specimens and subjected to qRT-PCR. Shown is a plot of the ratio of S/L. EIF1AX and PPA1 transcripts served as internal normalization controls. *P<0.05 using two-sided Student’s t-test. Variance was similar among groups being compared.