ColoSeq provides comprehensive lynch and polyposis syndrome mutational analysis using massively parallel sequencing

Abstract

Lynch syndrome (hereditary nonpolyposis colon cancer) and adenomatous polyposis syndromes frequently have overlapping clinical features. Current approaches for molecular genetic testing are often stepwise, taking a best-candidate gene approach with testing of additional genes if initial results are negative. We report a comprehensive assay called ColoSeq that detects all classes of mutations in Lynch and polyposis syndrome genes using targeted capture and massively parallel next-generation sequencing on the Illumina HiSeq2000 instrument. In blinded specimens and colon cancer cell lines with defined mutations, ColoSeq correctly identified 28/28 (100%) pathogenic mutations in MLH1, MSH2, MSH6, PMS2, EPCAM, APC, and MUTYH, including single nucleotide variants (SNVs), small insertions and deletions, and large copy number variants. There was 100% reproducibility of detection mutation between independent runs. The assay correctly identified 222 of 224 heterozygous SNVs (99.4%) in HapMap samples, demonstrating high sensitivity of calling all variants across each captured gene. Average coverage was greater than 320 reads per base pair when the maximum of 96 index samples with barcodes were pooled. In a specificity study of 19 control patients without cancer from different ethnic backgrounds, we did not find any pathogenic mutations but detected two variants of uncertain significance. ColoSeq offers a powerful, cost-effective means of genetic testing for Lynch and polyposis syndromes that eliminates the need for stepwise testing and multiple follow-up clinical visits.

Figure 1

ColoSeq workflow.

Figure 2

Depth of coverage. Depth of coverage from a representative patient sample run on a single lane of a HiSeq2000 instrument in a 96-plexed ColoSeq run. A: Distribution of coverage for each targeted base. Each bar represents the number of base pairs (in 1000s) at a particular depth of coverage. Of the targeted bases, 98.5% have a minimum of 50-fold coverage. B: Median coverage across each ColoSeq gene is shown (average coverage of 475-fold). Areas with depth of coverage >1000-fold reflect capture of highly homologous genomic regions.

Figure 3

Deletions and duplications detected by normalized depth of coverage. Three examples of genomic deletions (MLH1 exons 14-15, MSH2 exons 1-6, and PMS2 exon 8) and one example of a genomic duplication (MLH1 exons 6-12) detected by ColoSeq. Read depth is normalized across the 96 samples on a run. Significant deviations less than or greater than a normalized ratio of 1 across adjacent baited regions reflect genomic deletions (red) and duplications (blue). Genomic coordinates shown on the x axis are hg19.

Figure 4

Between-run and within-run reproducibility. Between-run and within-run reproducibility are shown as a function of minimum threshold for indel variants and calling single nucleotide (SNV). Variants were considered reproducible if they met or exceeded the specified cutoff value in the index sample and appeared in the replicate sample regardless of cutoff value. Reproducibility is shown for all variants in single-copy regions of exons and splice junctions. A threshold of 15 variant reads was chosen for the assay. Black squares, between-run; grey circles, within run.