Fig 1.
Somatic copy number alterations in 103 patients.
(A) Somatic copy number alterations in relapse free (n = 59) and progressed patients (n = 10) after radical prostatectomy at six months and (B) association with time to biochemical recurrence (the two groups were defined relative to the mean number of SCNAs in progressed patients (44 SCNAs) within the cohort). Progressed patients had significantly more SCNAs than relapse free patients (a mean of 19 compared to 44; Mann-Whitney U p = 0.0133). (C) Copy number burden in relapse free and progressed patients after radical prostatectomy at six months and (D) association with time to biochemical recurrence (the two groups were defined relative to the mean copy number burden in progressed patients within the cohort). Progressed patients had lower copy number burden (mean of 7.359 compared to a mean of 11.710; Mann-Whitney U p = 0.0166).
Fig 2.
Cluster dendogram and heatmap of somatic copy number alterations in 103 patients.
Cluster dendogram and heatmap of amplifications (red) and deletions (blue) representing the somatic copy number profile of the 103 patients calculated using unsupervised hierarchical clustering with Manhattan distance and complete inter-cluster linkage. Samples with SNP6.0 data are indicated with an asterisk. Confidence intervals determined by multiscale bootstrap resampling are displayed in S2 Fig.
Fig 3.
Cluster associations with clinical factors.
(A) Clusters association with Gleason Score (prostatectomies and TURP samples only). (B) Clusters association to risk of biochemical recurrence within C1 and C2-C5. At the time of writing the single prostatectomy case in C5 had a follow up time of 23 months and had not progressed.
Fig 4.
Frequency plot representing regions of amplification (red) and deletion (blue).
Across samples from (A) all 103 patients, (B) ETS-positive cancers, and (C) ETS-negative cancers. SCNAs present in more than one patient were plotted. The green lines represent the cut-off (5 patients) for defining recurrently altered regions. Chromosome numbers are indicated in black. Genes potentially involved in prostate cancer development in the minimal regions of alteration are annotated in black.
Table 1.
Genes that are recurrently altered by coding-changing mutations in regions of recurrent somatic copy number alteration defined by the extended MRAs.
Fig 5.
Examples of minimal recurrent deletions.
Deletions at (A) 6q14.3-q15, (B) 8p21.3-p21.2, (C) 10q23.31, (D) 13q14.13, (E) 14q24.1 and (F) 17q21.31. The genomic location of the MRA and linked genes are displayed. For the loss at 8p a recurrent region of homozygous loss (8:25417422–26386565) close to the MRA is indicated.
Fig 6.
Gains within the (A) q arm in Chr8, (B) focal amplifications at 7p11.2, and (C) 14p11.2. Three spatially separated regions of gain were present at 8q24.21. The chromosome gain events are represented with distinct colour blocks depending on the type of SCNA: gain (any gain in the number of normal allele copies) and amplification LOH (loss of one allele with any gain of the remaining allele). The regions of SCNA are ordered by length: top-smallest, bottom-largest. The genomic location of the MRAs and linked genes are additionally displayed.
Fig 7.
Minimal regions of deletion and amplification significantly associated with treatment failure.
(A), 1q42.2-q42.3 deletion, (B) 4p15.2-p15.1 deletion, and (C) 1q21.3-q22 amplification. Kaplan-Meier freedom from biochemical recurrence are displayed. Log-rank test probabilities adjusted for multiple testing by the Benjamini-Hochberg method are indicated. Ovals indicate the frequency of alteration.
Table 2.
Summary table of recurrent regions of deletion that follow the Knudson hit model.