Deciphering Complex Structural Variation with Linked-Reads

The association of chromosomal abnormalities with human disease was an important milestone in human genetics. Initial studies focused on direct observations of chromosomes using cytogenetic techniques. Newer methods such as aCGH have improved throughput and resolution, though at the cost of only detecting unbalanced SVs in targeted areas. Advances in high-throughput sequencing (HTS) show the potential to revolutionize the field. The use of HTS theoretically provides the ability to obtain information for both balanced and unbalanced events, but in practice this isn’t often achieved. Detection of large scale SVs is plagued by the fact that SVs are often mediated by breakpoints at repetitive sequences, for which standard short-read sequencing doesn’t provide unambiguous information. We have developed a new technology (Linked-Reads) that allows us to couple the power and accuracy of short read sequencing with long range information, providing the needed power for SV detection when compared to standard short read methods. We demonstrate the power of Linked-Reads to detect multiple types of structural variants, both balanced and unbalanced, in cancer and germline samples. In particular, we show identification of both balanced and unbalanced SVs larger than 30 Kb as well as power to detect substantially smaller deletion events. We demonstrate the power of Linked-Reads to identify translocations as well as more complex events involving multiple breakpoints. Our results on Linked-Reads demonstrate the potential to use a single assay and informatics system to perform molecular cytogenetics.