By Monica Heger
Researchers at the Mayo Clinic are currently evaluating different sequencing platforms and capture methods to use in molecular diagnostics, with the aim of having their first sequencing-based test ready within the next year.
Nicole Hoppman-Chaney, a fellow in clinical cytogenetics at the Mayo Clinic, told In Sequence that the researchers expect their first sequencing-based test to be for the diagnosis of hereditary colorectal cancer syndromes.
In a study published last month in Clinical Chemistry, the research team tested NimbleGen's capture array with both the Roche 454 GS FLX and Illumina Genome Analyzer. Additionally, they are assessing Agilent's in-solution capture technique, RainDance's microdroplet PCR enrichment technology, and standard PCR enrichment. Hoppman-Chaney said they would also like to evaluate Life Technologies' SOLiD and newer sequencing technology from Pacific Biosciences.
In the Clinical Chemistry paper, the team tested the techniques on five patient samples that had previously been tested for germline mutations in one of four colorectal cancer-related genes. They used NimbleGen's capture array to enrich exons in 22 genes, 21 of which have been associated with colorectal cancer, and then sequenced the exons on 454 and Illumina. They used Sanger sequencing to validate variants on a subset of the exons.
The researchers specifically chose samples with different types of cancer-causing mutations, including point mutations, small insertions, and small deletions, to evaluate the sequencing platforms' ability to detect different types of variants.
The targeted genes ranged in size from six to 106 exons, and a total of 410 exons were targeted by the array.
Of the variants validated by the Sanger method, six were missed by both 454 and Illumina, an additional six were missed just by 454, and two were missed by Illumina only. The variants that both sequencing methods missed were due to too few reads, which the authors attributed to a fault of the capture array, not the sequencing platform.
The variants missed by 454 only were frequently in homopolymer regions — a known issue with the 454 chemistry. "We know that if there is a mutation in that region we need to use a different method," Hoppman-Chaney said.
That is an issue that other groups have also found with the 454 platform. For example ARUP Laboratories' medical director for advanced technology, Karl Voelkerding, evaluated the 454 platform for use in hypertrophic cardiomyopathy diagnostics, and also found that it made errors in homopolymer regions (IS 3/30/2010). Both Voelkerding and Hoppman-Chaney said that while the newer Titanium chemistry now offered by Roche would improve the results, it would not completely eliminate the errors in those regions.
Additionally, the 454 platform detected 29 variants that were not validated by the Sanger method. The Illumina platform also did not detect these variants, and the researchers concluded that they were all false-positives. Only one of the variants was a known germline mutation, none were confirmed in the replicate samples, and most were found at low frequencies, the authors reported.
The main drawback of the Illumina platform is its shorter reads, which made detection of insertions and deletions difficult. In the study, the researchers obtained read lengths of only 36 bases. However, Hoppman-Chaney said that the longer reads that Illumina now offers would improve its ability to detect those mutations.
In terms of the NimbleGen capture array, she said its main problem is an issue inherent in all hybridization-based array capture methods — it captures a lot of "junk DNA," particularly among genes of high homology. "You'll capture what you want to capture, but you'll also capture DNA that's similar to your sequence if you have genes with high homology," she said.
As a result, the team is now also testing PCR-based enrichment. "It's more specific. You can design your primers so you only amplify what you want," she said. PCR enrichment has its own drawbacks, however. For example, sequence variants located beneath the primer can often cause allele dropout, she said.
Additionally, she said the group has been testing Agilent's in-solution capture method and RainDance's microdroplet PCR enrichment technology, but has not yet analyzed the results from those methods. The advantage of doing the capture in solution is that it would be less labor intensive and more high-throughput than array-based methods. She also said that other groups have reported good results from RainDance's technology, but that it would require the most upfront investment.
Hoppman-Chaney said the team is now evaluating data from the different capture technologies, and will then decide which combination of capture and sequencing platforms makes the most sense. She said the Mayo researchers plan to launch their first test in a year, but have not yet determined the price.
Despite some of the issues the Mayo researchers found with the sequencing platforms and capture techniques, Hoppman-Chaney said she thought next-gen sequencing would still have broad use in diagnostics. The next-gen sequencing platforms offer an "opportunity to sequence lots of genes at once," she said, which makes for a much faster turnaround time than Sanger sequencing.
Also, the technologies have improved since the Clinical Chemistry study was done, and are continuing to improve, she added. "These results were obtained a few years ago." The technologies "are just about ready."
Several labs and companies, such as Emory Genetics Laboratory, GeneDx, and Ambry Genetics, already offer sequencing-based genetic tests for diseases such as X-linked mental retardation, congenital muscular dystrophy, and hypertrophic cardiomyopathy (IS 5/25/2010).
Hoppman-Chaney said that she thinks there will eventually be targeted sequencing-based diagnostic tests for numerous diseases that will serve as a "bridge to sequencing the whole genome," which she said is still too expensive.