Myeloma immunoglobulin rearrangement and translocation detection through targeted capture sequencing

CluMP, clustering of mate pairs. ^Table S7.

Purple lines intersecting the x-axis indicate the thresholds chosen. (A) Clone counts for all clones plotted by clonal fraction. Blue indicates rearrangements manually verified to be true. Yellow indicates rearrangements manually verified to be false. Gray indicates rearrangements that were not examined manually. At a clonal fraction threshold above 0.1, all of the true rearrangements are captured while minimizing the number of false rearrangements that need to be manually verified. Plots (B, C) include only clones with a clonal fraction of >0.10. (B) Number of sequences plotted by the number of unique counts for each sequence (clone) in a cohort of 20 cell lines. All true rearranged clones had five or fewer instances of the clonal sequence present in the cohort of called rearrangements in 20 cell lines. (C) Number of clones plotted by absolute clone count of each clone. After filtering for clonal fraction >0.10, an absolute clone count captures all true rearrangements in this cohort without any false rearrangements.

Rearrangement in KMS11 Myeloma Cell Line and (B) IGH Rearrangement in MM1S. (A) Reads support a rearrangement between the two distinct loci (IGKJ5*01 and IGKV3D-15*01). Stacked, rainbow-colored reads indicate portions that map to the genomic locus in the corresponding split screen. The sequence (colored by nucleotide) of the multicolored reads in WGS data (top) matches that of the targeted sequencing data (bottom). The top track (KMS11 WGS) is publically available whole-genome data (Yang et al, 2014). The lower track (KMS11 targeted) is generated from the targeted sequencing probe set. Schematic shows V gene (green) and J gene (purple) with complementary sequences across multiple read pairs. (B) Reads support a rearrangement between the two distinct loci (IGHJ6*01 and IGHV3-30*01). Stacked, multicolored reads indicate portions of the read that do not map to reference genome; they comprise unique V-J junction sequences and continue on the corresponding split screen. The top track (MM1S_WGS) is publically available whole-genome data;³ the remaining tracks are WGS or targeted sequencing (targeted) data as indicated. MM1S and MM1R are steroid-sensitive and steroid-resistant derivatives of the same parent human myeloma cell line and bear the same IGHV-IGHJ rearrangement shown. Bioinformatics pipelines reliably call IGHV-IGHJ rearrangements in targeted data but not WGS data (see also Table S1). Manual verification demonstrates the presence of reads with identical sequences in WGS data as in targeted sequencing data.

,Breakpoints are indicated on relevant genes on chromosomes 4, 11, 16, and 20 with the IGH partner breakpoints in chromosome 14 indicated in corresponding colors. Gene locations are as indicated in the reference track. Genomic probe locations in IGH are as indicated. No probes were located on partner chromosomes.

Chr4 partner of t(4;14) translocation is displayed (chr14 not shown) for all six samples of the KMS11 dilution series: 1/10 dilution in the top panel down to 1/10⁶ in the lowest panel. Orange reads have mates that map to chromosome 14. Solid reads (orange or gray) map correctly to hg19 reference genome. Multicolored reads do not match reference genome at chr4 and map to the region on chr14. Mismatched reads are colored by base (A, green; T, red; C, blue; and G, orange). 1/10 and 1/10² dilutions are shown in compressed configuration, whereas more dilution samples are shown in expanded configuration for ease of manual review with fewer available reads. Three clear breakpoints are available to follow through the dilution series, “Br1,” “Br2,” and “Br3.” chr4 breakpoints are not baited in the targeted sequencing panel; reads mapping to chr4 displayed here are pulled down by probes targeting the chr14 breakpoint. Breakpoint 1 “Br1” has the greatest number of supporting reads to track through dilutions. The number of soft-clipped reads and mate pairs that support a translocation decreases as the sample becomes more dilute. Reads supporting the translocation at “Br1” are seen in the 1/10, 1/10², and 1/10³ dilutions but not in the 1/10⁴ dilution. Reads supporting the translocation at “Br3” are seen in the 1/10 and 1/10² dilutions, a single read supporting the translocation at “Br1” is seen in the 1/10⁵ dilution, and a single read supporting the translocation at “Br3” is seen at 1/10⁴ (circled in blue). Bases in this single read match those in the 1/10, 1/10², and 1/10³ dilutions, and the mate pair is found on chr14 at the correct corresponding breakpoint. Single reads illustrated at “Br3” in 1/10⁴ and “Br1” in 1/10⁵ dilution, respectively, are not definitive to identify a translocation in an isolated sample; however, the exact sequence and breakpoint location are highly specific to the cell line sample and can be used if there is prior knowledge (such as shown in less dilute samples).

In conventional library preparation, sample-specific indices were ligated to these dsDNA fragments directly, and minimal PCR amplification was performed. In barcoded library preparation, the end-repaired-A-tailed product underwent ligation to duplex barcodes overnight, which attaches unique tags to each molecule. Sample-specific indices were then added to each sample at the PCR step with a universal primer. After sequencing, samples can be identified and demultiplexed based on sample-specific index primers. Duplex barcodes can be disregarded if not proceeding with this type of analysis, and data are analyzed identically to non-barcoded libraries.