Comprehensive long-span paired-end-tag mapping reveals characteristic patterns of structural variations in epithelial cancer genomes

Abstract

Somatic genome rearrangements are thought to play important roles in cancer development. We optimized a long-span paired-end-tag (PET) sequencing approach using 10-Kb genomic DNA inserts to study human genome structural variations (SVs). The use of a 10-Kb insert size allows the identification of breakpoints within repetitive or homology-containing regions of a few kilobases in size and results in a higher physical coverage compared with small insert libraries with the same sequencing effort. We have applied this approach to comprehensively characterize the SVs of 15 cancer and two noncancer genomes and used a filtering approach to strongly enrich for somatic SVs in the cancer genomes. Our analyses revealed that most inversions, deletions, and insertions are germ-line SVs, whereas tandem duplications, unpaired inversions, interchromosomal translocations, and complex rearrangements are over-represented among somatic rearrangements in cancer genomes. We demonstrate that the quantitative and connective nature of DNA-PET data is precise in delineating the genealogy of complex rearrangement events, we observe signatures that are compatible with breakage-fusion-bridge cycles, and we discover that large duplications are among the initial rearrangements that trigger genome instability for extensive amplification in epithelial cancers.

Figure 1.

Structural variations (SVs) identified by dPET clusters of 15 cancer and two normal genomes. Column “Interpretation” indicates the genomic structure of the sequenced genome deduced from the mapping pattern of the dPET clusters to the human reference sequence (mapping to reference). Dark red arrows represent 5′ anchor regions and pink arrows represent 3′ anchor regions. Gray, blue, and red horizontal lines represent chromosomal segments. Red arrows indicate orientation of chromosomal segments. Asterisks indicate that clusters have been used for more than one insertion.

Figure 2.

Karyo-genomic maps of 15 cancer and two normal human genomes. Genomes are arranged in a circular manner with SV categories arranged in concentric layers as indicated on the top, left. Circular plots have been generated using Circos (Krzywinski et al. 2009).

Figure 3.

Comparison of SVs across 15 cancer and two normal genomes. (A–H) Frequencies (y-axis) of the indicated SV categories are shown for the individual genomes (x-axis). Cancer groups are separated by vertical gray lines. Degree of recurrent observation of the same SV is indicated in I, where 1 represents the observation in one genome and 17 represents the observation in all 17 genomes. (J) SVs that were observed in the normal individual(s) or which were observed in the cancer genomes, but match those observed in the normal individuals or match by >80% earlier described events (Korbel et al. 2007; Kidd et al. 2008) are indicated in dark blue. SVs that were also observed in the other 14 normal individuals are indicated in light blue. SVs observed only in cancer genomes are indicated in orange. The x-axis represents the number of genomes that share a particular SV, and the y-axis represents the frequency.

Figure 4.

Architecture and genealogy of amplifications in MCF-7. (A) Copy-number plots of chromosomes 1, 3, 17, and 20 with amplified regions (red boxes). (B) Concordant tag distributions are shown for amplified genomic regions (top, green track). Genomic segments between predicted breakpoints are indicated by colored arrows (middle) and dPET clusters with cluster sizes greater than 140 are represented by horizontal lines flanked by dark red and pink arrows indicating 5′ and 3′ anchor regions (bottom). Small to large dPET clusters are arranged from top to bottom. All but three dPET clusters were classified as complex. Mapping characteristics are described by: (Del) deletion; (IT) isolated translocation; (UI) unpaired inversion; (TD) tandem duplication. Cluster sizes are given for each cluster. (C) Possible genealogy of amplification. TD1,176 occurred early and subsequent rearrangements have pasted TD1,176 in different genomic contexts (G). (D–F) Double-color FISH using probes flanking TD1,176. Red, chr20:51,920,860–52,096,191; green, chr20:55,137,293–55,311,637. Double signals (filled arrowheads) indicate the fusion of the two loci and single signals indicate the normal genomic distance (open arrowhead). (D) Metaphase chromosomes, (E) metaphase nucleus, and (F) interphase nucleus showing amplification and fusion of breakpoint flanking sequences. (H) BMP7 (left) and ZNF217 (right) are juxtaposed by the TD1,176 rearrangement in a distance of 15,159 bp. (I) Models of local and interchromosomal amplification. Chromosomes are represented by gray and green horizontal lines. Amplified segment is represented by a red arrow. The initial tandem duplication (left) allows local amplification between two sister chromatids or homologous chromosomes (top) or interchromosomal translocation (bottom).

Figure 5.

The architecture of an amplification in primary breast tumor 14. (A) Concordant tag based copy-number estimate for chromosome 9 indicates an amplification of the distal region of 9p. (B) Concordant tag distribution of chromosome 9 position 2–10 Mb (top, green track). Genomic segments between predicted breakpoints are indicated by colored arrows (middle) and dPET clusters with cluster sizes greater than eight are represented by horizontal lines flanked by dark red and pink arrows (bottom). Abbreviations for mapping characteristics of dPET clusters are described in Figure 4. (C) Genomic structure of KDM4C. Location of amplified deletion (Del25) is indicated by dashed vertical lines. (D) Sequencing result of RT–PCR confirms the in-frame deletion transcript with the more upstream located exon 1.

Figure 6.

Accumulation of short-span unpaired inversions in amplified regions of gastric tumor 17. (A) PET mapping pattern of short-span unpaired inversions and the interpretation. The mapping of a 5′ anchor (dark red arrow) to the + strand and a 3′ anchor (pink arrow) to the – strand indicates a head-to-head fusion (red arrows) with increasing chromosomal coordinates closer to the breakpoint (top) and a 5′ – strand/3′ + strand mapping indicates a tail-to-tail fusion with decreasing chromosomal coordinates closer to the breakpoint (bottom). UI120 and UI118 in B are examples of head-to-head and tail-to-tail fusions, respectively. (B) Amplifications on chromosomes 5 and 18 of gastric tumor 17 are indicated by concordant tag counts (green). Cancer structural rearrangements with dPET cluster sizes >15 are indicated by dark red and pink arrows for 5′ and 3′ anchors, respectively. Abbreviations and figure structure are described in the legend of Figure 4B. Unpaired inversions with a breakpoint distance <40 Kb are indicated by asterisks. (C) Schematic representation of an isolated translocation between chromosome 5 (green) and 18 (gray). Black circles represent centromeres; blue X represents site of recombination; gray arrows indicate the direction of increasing genomic coordinates. (D) Interpretation of accumulated short unpaired inversions in amplifications by BFB cycles.