By Monica Heger
Illustrating one of the many quandaries of using next-gen sequencing in a clinical setting, researchers who were using exome sequencing to study attention deficit hyperactivity disorder may have found novel causative mutations for a patient's anemia — a scenario that raises the question of how to handle unrelated findings in sequencing studies.
The results of the study, conducted by researchers at the University of Utah, BGI, and the Children's Hospital of Philadelphia, were published in Discovery Medicine earlier this month.
The team set out to use exome sequencing in a pedigree — two parents and their four children. The father and the two sons were both affected with ADHD. The two daughters declined to participate in the study, so only the exomes of the parents and the two sons were sequenced. Over the course of the study, one of the sons, age 28, shared that as a child he had been diagnosed with idiopathic hemolytic anemia, meaning the cause of the anemia was never explained.
Since the researchers already had whole-exome sequencing data on the patient, they saw an opportunity to try and figure out the cause of the anemia.
Despite being a researcher as opposed to the patient's clinician, and a psychiatrist rather than a hematologist, "when the patient informed me he had anemia, and I had already sequenced his exome, I felt like I had an ethical obligation to try and figure out why he had anemia," Gholson Lyon, a research scientist at the Center for Applied Genomics at the Children's Hospital of Philadelphia and lead author of the study, told Clinical Sequencing News.
Analyzing the exome sequencing results of the affected son revealed that he had two very rare mutations in the PKLR gene, which encodes the pyruvate kinase, an enzyme that acts in glycolysis and disruptions of which are known to cause anemia. The patient was heterozygous, receiving one mutation from his mother and the other from his father. The mutations were confirmed with Sanger sequencing.
The team also obtained permission to examine the patient's medical records, which revealed that he was deficient in pyruvate kinase, supporting the evidence that the PLKR mutations were the cause of the patient's anemia. The findings were also supported by biochemical analyses.
Despite the support from the patient's own medical records as well as biochemical analyses, the sequencing was not done in a CLIA lab, nor were the results validated with a certified test, so they could not be returned directly to the patient.
Instead, Lyon called the patient's hematologist and informed her that he had likely found the cause of the patient's anemia, but that the results had not been confirmed in a certified setting.
"I have this finding that I got in a non-CLIA lab, that's good enough for research, but not good enough to give back to the patient, so what do you do? That's the crux of the problem," Lyon said.
While there are no clear guidelines on handling such findings from regulatory agencies or other governing bodies, several groups and genome centers have begun addressing the issue on their own. For instance, the International Cancer Genome Consortium's policy on returning results is to leave the decision to the local project leaders, since each country has a different legal environment; however, the consortium is now reviewing the practices of the different groups, and will consider changing its policies (see story, this issue).
[ pagebreak ]
Additionally, the Genome Institute at Washington University has its own framework in place for returning results to a patient's physician, while still keeping the patient's anonymity. This spring, those guidelines were put to the test on two separate occasions (CSN 4/26/2011).
In one case, whole-genome sequencing was done with the intent to diagnose the patient, and medically relevant results were returned to the patient's physician and used to guide her course of treatment. In another case, the whole-genome sequencing of a deceased patient's tumor uncovered de novo TP53 mutations, which will be relevant to the patient's children. In both cases, the institute's "moveable firewall protocol" worked as intended, allowing the clinicians to be informed of the results and to explain them to the patient or the patient's family.
While there have only been a few documented cases of finding either unrelated or unintended medically relevant results, as whole-genome and whole-exome sequencing becomes more widespread, these situations will occur more often.
"I think the solution going forward is that we need to sequence everyone's genomes in a CLIA-certified laboratory," said Lyon. "It makes no sense to sequence the genome of someone in a non-CLIA certified lab and then grapple with what to do with these unrelated findings."
While Lyon believes he did the ethical thing, he acknowledged that some might criticize him, saying that he should not have looked for the patient's anemia cause in the first place, or that he should have verified the results in a CLIA-certified lab before informing the hematologist.
"My response is that I'm trying to understand ADHD, and I'm not focused on anemia. It was an unrelated finding that came up in the course of my research. I don't have the resources to develop a CLIA-certified test and prove it in a CLIA-certified way," he said.
In terms of understanding ADHD, he said that the exome sequencing yielded some interesting mutations that may be involved in the disease, and that he is now testing those variants further.
The exomes of the three family members were sequenced using Agilent's SureSelect in-solution technology and sequenced on the Illumina Genome Analyzer at BGI. The team looked for rare variants with a minor allele frequency of less than 1 percent that were present in the father in sons, but not the mother and daughters, assuming that causative common variants would likely have been found already from genome-wide association studies.
Three different pipelines were used to filter variants — SOAP, GATK, and SNVer. Additionally, the team compared results to around 6,000 other exomes that had been sequenced for various projects by teams at BGI and Baylor. The researchers identified four plausible gene candidates containing rare, nonsynonymous variants: ATP7B, CSTF2T, ALDH1L1, and METTL3. All four genes have also been shown to be expressed in the brain and have been implicated in other neuropsychiatric conditions.
Have topics you'd like to see covered by Clinical Sequencing News? Contact the editor at mheger [at] genomeweb [.] com.