November 30, 2011 in GenomeWeb – Clinical Sequencing News
By Julia Karow <mailto:jkarow@genomeweb.com>

Baylor College of Medicine's Cancer Genetics Laboratory is about to launch a sequencing-based test on the Ion Torrent PGM that uses the firm's Ion AmpliSeq Cancer Panel to target mutations in 46 cancer genes.
The lab is among the first in the country to use targeted next-gen sequencing in a clinical setting for cancer. Washington University School of Medicine's Genomics and Pathology Services Laboratory recently introduced a similar test on the Illumina HiSeq, targeting 28 cancer genes (CSN 11/22/2011 <http://www.genomeweb.com/sequencing/wash-u-med-school-offers-28-gene-cancer-dx-panel-hiseq-through-clia-lab> ).
Baylor's test, which the CLIA- and CAP-certified CGL will start offering in December, targets 739 mutations in 46 commonly mutated cancer genes. It will be priced under $2,000, and its turnaround time will be around seven to 10 days, although it is possible to complete the test within a day or two, according to Marilyn Li, the lab's director and a professor of molecular and human genetics at Baylor.

While Baylor will offer the test on a research basis, it can be ordered by both basic researchers and doctors. "It's truly a research tool at this point in time," said Condie Carmack, the lab's general manager. And while insurance will not initially pay for it, the CGL will look into whether the test could be reimbursable. The panel could replace, at a lower cost, existing tests that sequence small numbers of genes, he said.
Initially, the test will be based on the Ion AmpliSeq Cancer Panel that Ion Torrent launched for the PGM in October (CSN 10/12/2011 <http://www.genomeweb.com/sequencing/life-tech-launches-cancer-panel-amplicon-sequencing-kit-ion-pgm-plans-510k-filin> ). That panel uses single-tube PCR to amplify 190 amplicons in 46 cancer genes from 10 nanograms of DNA in 3.5 hours, according to the company, and can detect mutations down to a frequency of 5 percent.

According to Li, the coverage for the targeted mutations varies between less than 100x to close to 10,000x, with an average coverage of about 2,000x.So far, the CGL has not had problems with homopolymer regions, which the PGM is said to have trouble with. "We are aware that it could be a problem [but] it does not seem to be an issue with this AmpliSeq product," Carmack said.
Ion Torrent selected the genes for the panel after seeking input from several cancer researchers, including the Baylor team. It includes somatic mutations in genes commonly mutated in cancer as well as germline mutations found in inherited forms of the disease.The reason CGL chose the Ion Torrent platform for the test is its quick turnaround time and low cost per run. At the moment, the lab has two Ion Torrent sequencers that share one Ion Server, but it considering purchasing "a couple" more PGMs.

 
The Illumina MiSeq did not exist at the time they were developing the test, according to Carmack, and while the lab also has a HiSeq, its turnaround time was too slow and its throughput too large. "We really wanted something quick, inexpensive, and compact," he said.
Another advantage of the Ion Torrent panel is that it can be customized, and in future versions of the test, the CGL plans to add or omit genes or to sequence some genes with greater depth. "That is very appealing to a lot of researchers, as well as to clinical trials," Li said. According to Ion Torrent, AmpliSeq custom panels can include "hundreds of genes" or DNA regions up to 500 kilobases in size.
Initially, the test will run on the Ion 314 chip, though CGL is considering moving it over to the higher-throughput 316 or 318 chips. This would allow them to deepen the coverage, add more genes to the panel, or multiplex samples.
While the lab is currently not using the Ion OneTouch to prepare samples for the PGM, it already has the instrument in house and is planning to test it.

Focusing on the 'Clinically Useful'
There are several reason why CGL opted for a targeted cancer gene panel, rather than analyzing whole cancer exomes or genomes, Li said.
For one, the mutations targeted by the panel are interpretable, whereas large-scale sequencing yields many variants that "may not be clinically useful," and confirming them all takes a long time. "That doesn't meet the need of cancer diagnosis," she said. In addition, the cost of targeted sequencing is still much lower than that of whole-exome or whole-genome sequencing, and the turnaround time is shorter.
Also, because the test is less complex than exome or genome sequencing, it was easier to implement it in the clinical lab, she said. Lab technicians were able to achieve the same results as the R&D lab relatively easily.
For whole-exome or whole-genome sequencing for cancer to become practical in a clinical setting, she said, its cost will have to come down to that of a single-gene test, and the data analysis will need to be able to extract relevant results quickly.

Another Baylor lab, however, is already forging ahead with clinical whole-exome sequencing, albeit for inherited diseases rather than cancer. Earlier this month, Baylor announced the opening of the CLIA-certified Whole Genome Laboratory, which offers a whole-exome sequencing test for the diagnosis of genetic disorders, with plans to move to whole-genome sequencing in the future (IS 11/16/2011 <http://www.genomeweb.com/sequencing/baylor-whole-genome-laboratory-launches-clinical-exome-sequencing-test> ). Like the CGL, the WGL is affiliated with Baylor's department of molecular and human genetics and its Medical Genetics Laboratory, but it also closely collaborates with the Human Genome Sequencing Center.

There are some types of mutations that the CGL's current cancer panel, which focuses on point mutations and small insertions and deletions, cannot detect — for example, translocations, copy number variants, or epigenetic changes. "There is no one technology that is going to cover it all," Li said. But CGL is hopeful that further improvements of the Ion Torrent platform will enable it to analyze both copy number variants and translocations.
The cancer panel is only one of more than 135 tests the CGL offers, including sequencing of single genes, deletion and duplication testing, chromosome analysis, FISH, and chromosomal microarray analysis.
Many of these tests, which target mutations found in inherited cancers, were already offered by Baylor's Medical Genetics Laboratory, from which the CGL split off about a year ago. "We are moving them over under the CGL banner and expanding our content in somatic and acquired cancers," Carmack said. The reason CGL was founded as a separate entity is that its focus on both non-inherited and inherited forms of cancer required special expertise in cancer genetics, he explained.

The CGL — a joint project of the department of molecular and human genetics, the department of medicine's hematology and oncology division, the department of pathology, the Dan L. Duncan Cancer Center at BCM, and the Texas Children's Hospital Pathology Laboratory — still works closely with the MGL, which shares the same location, about a mile and a half from the main Baylor campus.
Besides molecular testing for guiding patient treatment and prognosis, the CGL offers clinical trial services to companies and conducts research to discover new cancer genes, markers, and tests for them, both on its own and in collaboration with industry.