SAN FRANCISCO (GenomeWeb) – Researchers at Stanford University and Intermountain Healthcare have used 10x Genomics' linked-read technology to tease apart complex rearrangements in gastric cancer metastases from a single patient that could not be parsed by conventional sequencing.
Lincoln Nadauld, executive director of precision medicine and precision genomics at Intermountain, said that the proof-of-principal study, published recently in Genome Medicine, highlights the potential benefits of evaluating cancer genomes with long-range sequence information. For instance, he said, the alterations identified included the amplification of a known gastric cancer gene, FGFR2, for which there are treatment options.
"It serves as an example of the broader applications for this technology — to analyze living patients' tumors in search of complex large structural events that could change the way they're treated," he said.
The researchers are now planning additional studies using the 10x Genomics technology, including to evaluate other cancer patients' tumors. In addition, he said, they are currently using it to evaluate a family whose clinical symptoms and outcomes hint at inherited cancer, but for whom sequencing has failed to uncover pathogenic variants. The genomics team at Intermountain is also collaborating with the cardiology team to evaluate banked samples from patients with extreme cardiovascular disease phenotypes to see if structural variants could explain the phenotype, Nadauld said.
Ultimately, if the researchers demonstrate that the technology has clinical utility and is economically feasible, they will look to incorporate it into their existing clinical workflow. Intermountain Healthcare has been offering a targeted sequencing test known as ICG100 for several years now for late-stage cancer patients. It has also developed a gene fusion panel, another targeted sequencing pan-cancer panel, an acute myeloid leukemia panel, and a whole-genome sequencing protocol, all of which are run in various research projects.
For the Genome Medicine study, Nadauld collaborated with researchers at Stanford University, led by Hanlee Ji's laboratory there. Stanford and Intermountain forged an agreement last year to collaborate on clinical genomics research, under which researchers at the Stanford Genome Technology Center would work with Intermountain to bring cutting-edge genomics technology into translational and clinical research.
In the study, the researchers decided to analyze somatic alterations in samples from a patient with gastric cancer who had developed metastases in her ovaries. Nadauld said they chose this case because the cancer was so "unusual," causing the researchers to suspect that there was a complex structural variant.
"Identifying cancer rearrangements is one of the most daunting tasks we have," Ji said. "A lot of the issues with using genomic sequencing approaches stem from the fact that in standard preparations, you lose the genomic contiguity that would help to detect those rearrangements," he said.
Ji added that the group decided to use the 10x technology both because it would enable analysis of large DNA molecules around 50 kilobases in size and because the DNA input requirements were small enough to be feasible for clinical cancer samples. Only around 1 nanogram of DNA is needed, he said, which "enables analyses from biopsy samples."
The researchers performed linked-read whole-genome sequencing of both metastatic tumors and matched normal tissue. The DNA from the primary tumor was too low in quality to perform linked-read sequencing, but the researchers performed conventional sequencing of the primary tumor as well as the metastatic tumors.
The researchers identified an amplification of the region surrounding the FGFR2 gene in both metastatic samples but not in the primary tumor, findings that were confirmed by ddPCR.
While conventional whole-genome sequencing was able to identify some of the structural variant events, it could not identify all of them, and did not provide enough detail to reconstruct the series of events that led to the FGFR2 amplification.
By contrast, the linked-read data yielded more granular details of the amplification, including structural variant events that were unique to each metastatic tumor, and enabled the researchers to phase the variants and piece together the processes that led to the rearrangement.
The analysis identified that each metastatic tumor had different breakpoints, suggesting that the rearrangements occurred independently. The researchers were also able to phase the structural variant events using the linked-read data. In one metastatic tumor, they evaluated a duplication, a deletion, and an inversion, and found that they all occurred on the same haplotype.
For the second metastatic tumor, linked-read data uncovered five distinct structural variants — two duplications, two deletions, and one inversion — and again found that they were all on the same haplotype.
The researchers next validated the rearrangement by assembling the FGFR2 region de novo. They also performed functional studies, confirming the role of FGFR2 as a driver in the patient's metastatic disease.
"We realized that what would have been considered to be a normal amplification by conventional whole-genome sequencing was, when we drove down into the structure, a whole series of complicated events," Ji said.
Those findings could have implications for therapy or prognosis. For instance, he said, the context in which an amplification exists can affect whether or not a patient responds to a certain targeted therapy.
Ji said that his team is now using the 10x technology to evaluate other gastro-intestinal malignancies and is "seeing very consistently events that would not have been recognized" by other methods.
In the future, he envisions that it would be possible to use the technology in a clinical setting. For cancer applications, he said, one possibility is that the technology could replace traditional cytogenetic or fluorescent in situ hybridization assays, which test for specific structural variants in leukemia. The advantage of replacing FISH-based tests with 10x technology is that it would increase the resolution and provide both structural variant information as well as point mutations. And, because linked-read sequencing would provide genome-wide information about structural variants, it would enable researchers to assess both known variants with predictive and prognostic relevance and also the general stability of the tumor, Ji said.
Nadauld added that in order to use the 10x Genomics technology in a clinical workflow, studies would have to be done that demonstrated both its clinical utility and also cost effectiveness, similar to what the group has demonstrated for ICG100 panel.