SAN FRANCISCO — Increasingly, issues related to sample prep, not sequencing technology, are becoming the hurdle for developing clinical next-gen sequencing-based tests, researchers from Quest Diagnostics and the US Centers for Disease Control and Prevention said at this week's Molecular Medicine Tri Conference in San Francisco.
"A lot of the importance and value of a next-gen sequencing diagnostic test is in the sample prep," Jamie Platt, director of advanced sequencing at Quest Diagnostics, said during a presentation here.
For instance, said Platt, many clinical NGS tests attempt to take advantage of next-gen sequencing's greater resolution compared to Sanger sequencing by looking for mutations at ever-lower frequencies. Sanger sequencing cannot detect variants below about 20 percent frequency, while next-gen sequencing can detect variants present at around 1 percent frequency and sometimes even lower. However, said Platt, detecting such low-frequency variants is dependent on the sample prep.
She described an example of an HIV sample that contains 200 copies of the virus per mL of blood. If a mutation is present at only 1 percent frequency, there will only be two copies of the mutation per mL. Clinicians typically draw around 500 microliters of blood, and then elute it in 50 mL, reducing the number of HIV molecules to 100, with only one mutant in the sample.
Sample prep, including PCR amplification, reduces volume even further. "So you may not have any of your mutant in these reactions," she said. "If the mutant is not in your sample, no technology will get you the answer."
Quest has developed an HIV tropism test that uses a combination of Sanger sequencing and next-gen sequencing on Roche's 454 GS Junior platform. The test determines whether patients are eligible to receive the drug maroviroc, marketed by Pfizer as Selzentry.
A tropism test is required before receiving Selzentry because the drug is effective only in patients harboring so-called R5 viruses, which use the CCR5 chemokine coreceptor to infect cells. Selzentry binds to CCR5, so is not effective against X4 viruses, which enter the immune cells through the CXCR4 coreceptor.
Quest's test first does Sanger sequencing as a quick method to identify X4. If a patient is X4 based on the Sanger test, that can be reported out quickly, Platt said. However, since the Sanger test can only identify X4 if it is present at 20 percent frequency or above, the cases that are predicted to be R5 reflex to the NGS assay.
Platt said that in order to ensure low frequency variants are selected for, Quest samples in triplicate, running three PCR reactions from the same blood sample. "It's all about the sensitivity in the beginning," she said.
Moving forward, Platt said that it may become necessary to draw more blood to get at minor mutants. Additionally, all steps in the sample prep process must be carefully evaluated, including nucleic acid extraction and library preparation methods.
Often, nucleic acid extraction methods with the highest yield and purity do not recover all the targets, she said. Additionally, nucleic acid extraction can introduce other biases, such as methylation, that may be magnified downstream.
Next-gen sequencing library prep methods also have the tendency to introduce bias, Platt said. Hybridization-based methods may capture off targets. For amplicon-based methods, targets must be precisely defined.
Aside from its HIV tropism test, Quest has also developed a next-gen sequencing test to diagnose Charcot-Marie-Tooth disease, and it recently struck multi-year deals with both Illumina and Life Technologies to use their respective sequencing technologies for clinical testing.
Martin Siaw, associate scientific director for advanced sequencing at Quest Diagnostics, said that the company is platform agnostic and is not only testing out the various next-gen sequencing platforms, but also how easily sample and library prep on the platforms can be automated.
Patricia Mueller, chief of the Molecular Risk Assessment Laboratory in the Newborn Screening and Molecular Biology Branch of the CDC, highlighted yet another issue with sample and library prep for NGS tests — pseudogenes.
Mueller said that her lab has attempted to validate an NGS screening test for over 400 recessive diseases in newborns, including congenital adrenal hyperplasia. The disease is caused by a mutation to the CYP21A2 gene, but the gene also has a non-functional pseudogene — CYP21AP. The gene and pseudogene are 98 percent homologous, she said. And while the two can be distinguished by generating a long-range amplicon, the amplicon must span 5,000 base pairs, which is not possible with short-read sequencing technology, she said.
When testing the newborn screening panel, Mueller said that the team tried both hybrid capture arrays and amplicon-based target capture steps, but "both enrichment approaches failed to separate the functional gene from the pseudogene."
The problem is significant, Mueller added. Studies have identified between 12,000 to 20,000 regions that are likely to encode for pseudogenes. Parent genes have been identified for more than 9,000 pseudogenes, including more than 3,000 known functional genes with pseudogenes.
Going forward, Mueller said that pseudogenes need to be identified and catalogued so that next-gen sequencing test developers are aware of them. Additionally, better methods will need to be developed to ensure that target enrichment steps select for the functional gene, not the pseudogene.