Figure 4

In silico and in situ validation. (A) Coverage of overlapping aberrant and non-aberrant reads at the REARR breakpoints of clonal TMPRSS2-ERG REARRs in two prostate adenocarcinomas (cases P03-2345 and PR-09-146). Upon removal of admixed DNA reads (gray), local coverage of aberrant and non-aberrant reads match (AP close to 0.5), supporting the clonality of the rearrangement. (B) Representative case (PR-2525) with MSR1 subclonality REARR, validation by FISH. Low power view of prostate adenocarcinoma, Gleason score 3 + 4 = 7 in a prostatectomy specimen (black box). Some areas do not have deletion of MSR1 as demonstrated by the presence of two yellow signals in tumor nuclei (yellow box). Other areas show hemizygous deletion of MSR1 as demonstrated by the presence of only one yellow signal in tumor nuclei (blue box). Occasional nuclei with wild-type MSR1 (arrow heads) are identified in this area. Nuclei with MSR1 hemizygous deletion comprised approximately 30% of assessed tumor areas. (C) A homozygous subclonal deletion including CHD1 is inserted within a large hemizygous clonal deletion (case PR-2741). Cancer AF highlights the different proportion of aberrant reads in the two cases. (D) Representative case with CHD1 subclonality (case PR-2741), validation by FISH. Low power view of prostate adenocarcinoma, Gleason score 4 + 3 = 7 in a prostatectomy specimen. Some areas have homozygous deletion of CHD1 as demonstrated by the presence of only two green signals (reference probe) in tumor nuclei (bottom left blue box). In contrast, other areas show hemizygous deletion of CHD1 as demonstrated by the presence of one red (CHD1) and two green signals (reference probe) in tumor nuclei (bottom right black box). Note the presence of two red and two green signals (wild-type CHD1) in adjacent stromal cells, used as internal control (arrow heads).