Desktop transcriptome sequencing from archival tissue to identify clinically relevant translocations

Abstract

Somatic mutations, often translocations or single nucleotide variations, are pathognomonic for certain types of cancers and are increasingly of clinical importance for diagnosis and prediction of response to therapy. Conventional clinical assays only evaluate 1 mutation at a time, and targeted tests are often constrained to identify only the most common mutations. Genome-wide or transcriptome-wide high-throughput sequencing (HTS) of clinical samples offers an opportunity to evaluate for all clinically significant mutations with a single test. Recently a "desktop version" of HTS has become available, but most of the experience to date is based on data obtained from high-quality DNA from frozen specimens. In this study, we demonstrate, as a proof of principle, that translocations in sarcomas can be diagnosed from formalin-fixed paraffin-embedded (FFPE) tissue with desktop HTS. Using the first generation MiSeq platform, full transcriptome sequencing was performed on FFPE material from archival blocks of 3 synovial sarcomas, 3 myxoid liposarcomas, 2 Ewing sarcomas, and 1 clear cell sarcoma. Mapping the reads to the "sarcomatome" (all known 83 genes involved in translocations and mutations in sarcoma) and using a novel algorithm for ranking fusion candidates, the pathognomonic fusions and the exact breakpoints were identified in all cases of synovial sarcoma, myxoid liposarcoma, and clear cell sarcoma. The Ewing sarcoma fusion gene was detectable in FFPE material only with a sequencing platform that generates greater sequencing depth. The results show that a single transcriptome HTS assay, from FFPE, has the potential to replace conventional molecular diagnostic techniques for the evaluation of clinically relevant mutations in cancer.

FIGURE 1

Work flow for analysis. See the Materials and methods section.

FIGURE 2

Split screen for fusion spanning pairs of SS18 and SS1 in sample SS_16. Gray reads represent paired reads that both map to the same transcript, either SS18 (arrows A) or SSX1. The black reads in SS18 window are paired with the tan reads in SSX1 window (arrows B). Notably, there are no reads 5′ to the breakpoint in SSX1, which suggests that the unaffected allele is not expressed. There is some expression of the unaffected SS18 as demonstrated by some reads continuing to map 3′ to the breakpoint. The read coverage is displayed above.