Researchers at the Radboud University Nijmegen Medical Center are looking at ways to expand their diagnostic exome sequencing program for a range of diseases, even as they go through data from the pilot phase of the effort.
"We will continue to offer exome sequencing in the diagnostic setting," Radboud University Nijmegen human genetics researcher Joris Veltman told Clinical Sequencing News.
He explained that the extent to which the program will expand is "more of a financial discussion at the moment than a discussion of whether the test is good or not."
Nevertheless, the Nijmegen researchers hope to provide diagnostic exome sequencing to between 1,000 and 2,000 patients in the coming year, focusing on conditions that may stem from mutations to many different genes.
"Everybody [involved] has agreed that for the diseases we've identified as genetically heterogeneous diseases, exome sequencing is a major step forward," he said.
Veltman introduced the diagnostic exome sequencing pilot project at the joint International Congress of Human Genetics/American Society of Human Genetics meeting in Montreal last fall (CSN 10/19/2011).
For the program's pilot stage, investigators at Radboud University Nijmegen Medical Center set out to do diagnostic exome sequencing for around 500 patients, focusing on heterogeneous genetic conditions such as blindness, movement disorders, and intellectual disability that were recalcitrant to diagnoses by conventional means.
At the European Society of Human Genetics meeting in Germany this summer, Radboud University Medical Center's Marcel Nelen presented data on more than 450 of the exomes analyzed as part of that effort (CSN 7/3/2012).
In a study appearing online today in the New England Journal of Medicine, meanwhile, Veltman and his colleagues reported on results from pilot efforts to use exome sequencing on parent-child trios as a strategy for diagnosing severe, sporadic cases of intellectual disability that could not be explained by other genetic tests.
Of the 100 children tested for the study, the researchers found causal mutations in 16 cases. Most of these mutations were de novo changes not present in parents' coding sequences, though the team also identified three cases of intellectual disability stemming from X-linked mutations inherited from the mother.
Ten of the 13 de novo mutations and all three X-linked mutations fell in genes implicated in intellectual disability in the past. But analyses of the de novo alterations also led to the identification of three new intellectual disability risk genes that were subsequently found to be mutated in at least one other affected individual screened during follow-up testing.
Though the diagnostic yield of 16 percent is a bit lower than that reported for some of the other conditions being tested by exome sequencing at the Radboud University Nijmegen Medical Center, it is on par with the percentage of intellectual disability cases that can be diagnosed using array-based methods that look for suspicious duplications and deletions.
"The results of the study definitely show that exome sequencing is a very important diagnostic test for this patient group, so we will continue to expand that," Veltman said.
He noted that the group is currently re-analyzing the exome sequence data to determine whether they may have under-called de novo mutations in the current analysis, since the burden of de novo mutations per affected individual was a tad lower than that reported in other studies of autism, intellectual disability, or neurological conditions.
It's also possible that the exomes will have to be sequenced to deeper coverage to pick up all of the de novo mutations, he explained. Consequently, the Nijmegen team is trying to find the balance between sequencing to sufficient coverage while still testing as many patients as possible.
"The [number] of de novo mutations is a little on the low side and we have to improve that by improving the data analysis as well as by sequencing at greater depth," Veltman said.
The researchers expect to further increase the number of cases that can be diagnosed by exome sequencing as new gene culprits are verified in individuals with well-characterized symptoms.
As the Nijmegen researchers expand the diagnostic exome sequencing program over the coming year, they will focus on conditions that may stem from mutations to many different genes.
Cases of severe, sporadic intellectual disability that are not associated with a genetic syndrome are high on that list of priority conditions, Veltman explained, because they can involve hundreds of known and yet-undiscovered genes, often making them tricky to diagnose.
Moreover, mutations in these genes tend to be rare, with a series of several recent studies indicating that de novo mutations are over-represented in the genomes of those with intellectual disability, autism, and other neurological conditions (see, for example: IS 11/16/2010, CSN 9/28/2011).
In a study published online in the Lancet last week, for instance, German and Swiss researchers used exome sequencing to look for de novo mutations in 51 individuals with severe sporadic, non-syndromic intellectual disability. That analysis indicated that between 45 percent and 55 percent of such cases could be caused by mutations or small insertions and deletions that spring up de novo.
Because of the combined rarity and heterogeneity of mutations involved in sporadic intellectual disability, such cases often remain undiagnosed even after several rounds of genetic and other diagnostic testing.
With that in mind, the Dutch team speculated that exome sequencing could be a cost-effective way of coming up with diagnoses for a significant fraction of these individuals, while at once identifying new causal genes for intellectual disability to help diagnose others in the future.
"One of the advantages of using this technology for intellectual disability is that there are so many different genes that can cause the disorder," Heather Mefford, a physician and researcher in pediatrics and clinical genetics at the University of Washington, told CSN.
Mefford was not involved in the new NEJM study, but wrote a related commentary article for the same issue of the journal.
"I think the disadvantage right now of exome sequencing in a clinical setting is that we still don't always know when a change or a mutation is truly a mutation," she said. "If it's a gene that has not yet been implicated in the disease then you really do need to be sure that it's the cause."
She credits the Dutch team with doing "a really nice job" of verifying new causal mutations in the current study, but cautioned that such analyses are tricky and that extensive candidate gene validation may not be feasible for some diagnostic labs.
For the newly published analyses, the Nijmegen team tested 53 females and 47 males with non-syndromic intellectual disability.
Given the apparent importance of de novo mutations in this condition, diagnostic testing for that arm of the Nijmegen pilot program involved exome sequencing on affected individuals as well as their parents.
Each family received genetic counseling to discuss incidental findings and other considerations related to the use of exome sequencing for diagnosis.
The researchers captured coding sequences from each individual's genome with the Agilent SureSelect kit and sequenced them on the SOLiD 4 platform. On average, the sequences generated covered some 87 percent of the coding sequences targeted, with a median depth of around 64-fold per exome.
To track down mutations that might be relevant to disease in each patient exome, the team first turned to methods developed in the research realm, though candidate mutations were put through several rounds of validation before they were reported as causal.
"In this project we really made a switch from research to diagnostics," Veltman said.
"If you do this in a diagnostic setting it means that you're actually going to give a diagnostic report for each patient," he added. "So you have to be, of course, very sure that what you report back as causal is indeed causal."
All told, the group found 79 de novo mutations in the exomes of 53 individuals with intellectual disability.
For the potentially pathogenic mutations — those predicted to produce changes in the amino acid sequence encoded by a given gene — the team then looked at the identity of the gene affected, focusing on genes previously associated with intellectual disability.
When the mutation did appear in a known intellectual disability gene, the researchers compared the phenotype of the patient in question with those described for others with mutations in that gene.
In situations where the suspicious mutation occurred in a new gene, they considered everything from the gene's functional role to the inheritance model association with the mutation of interest before taking a smaller set of candidate genes forward to functional experiments and validation testing.
For the five strongest candidate genes that came out of those analyses, the investigators looked for recurrent ties to intellectual disability by doing targeted resequencing of the genes in 765 more individuals with intellectual disability.
At the end of these analyses, the team found 16 mutations that were deemed causal in the 100 intellectual disability cases considered.
Thirteen were de novo mutations — 10 affecting genes with known roles in intellectual disability and three found in newly identified risk genes.
Three more cases seemed to stem from maternally inherited, X-linked mutations, the researchers reported, and none of the causal mutations identified were inherited in an autosomal recessive manner.
Veltman noted that de novo mutations were also detected in 19 plausible candidate genes. But because the group wanted to be stringent and conservative in the results returned to patients and their doctors, these will only be reported as causal if and when they are further validated and found in additional intellectual disability cases.
Though the group has not settled on a standard timeframe for revisiting unsolved cases, Veltman said it is in the process of re-analyzing exome sequences from some individuals tested in the pilot program who have not yet received a diagnosis.
In her commentary, Mefford agreed that exome sequence-based diagnostic testing will likely require somewhat regular re-analyses to incorporate information on new risk genes found in the future.
She also highlighted the importance of carefully defining causal mutations to such genes as a means of making exome-based diagnoses more straightforward, particularly because many diagnostic labs may not have the time and/or resources to test candidate mutations functionally or through targeted sequencing-based screening studies.
"Most diagnostic labs will neither have the samples from patients nor the resources to follow up every putative mutation affecting a gene in this fashion," she wrote, "so there must be continued collaboration between researchers and clinicians."
"A systematic approach to sharing data on sequence variants among laboratories will facilitate the validation process," she added, "because large numbers of patients are required to determine whether a variant of undetermined significance is in fact causal."
The amount of validation testing needed to verify authentic causal mutations within the pool of candidate glitches is expected to dip somewhat as the list of genes that can contribute to intellectual disability and other heterogeneous conditions grows.
"I think we're going to accumulate data really quickly, both from research studies and from clinical studies," Mefford said. "And as long as that information is shared in an efficient way, these types of large validation studies will become less and less necessary."
Veltman, too, is keen to get a better idea of which genes carry recurrent mutations in each of the conditions his group is testing by exome sequencing, since that knowledge is also expected to bump up the diagnostic yield of such tests.
For the pilot phase of the Nijmegen program, insurance companies in the Netherlands reimbursed the sequencing-based diagnostic tests at a rate of around €1,500 ($1,940) apiece.
Veltman said that going forward, the reimbursement rate is expected to be even lower — closer to €750 ($970) per exome sequenced. And that is prompting researchers to look for ways to rein in costs at each step.
"The costs for exome sequencing really have to go down for us as a laboratory to offer this on a larger scale," he said.
For instance, Veltman said the group is in discussions with Life Technologies to determine whether SOLiD sequencing is an economically viable option as the program continues to expand.
"We are now expanding our exome sequencing into different disorders — for example to immune deficiencies," Veltman noted. "We could easily expand this to many more diseases, so that's why we need to get the cost down and the throughput up."
But even though exome sequencing has a higher cost than other genetic tests, including single gene tests, Veltman argued that it could still be a cost-effective option for tier 1 diagnostic testing of certain conditions, since many children currently undergo several rounds of testing without getting a firm diagnosis.
In addition to the types of analyses already being done with whole-exome sequence data, he added, the group is starting to use the datasets for copy number variation analyses as well.
"We hope that this test, which is of course picking up much more variation, within one or two years will become the standard test for intellectual disability," Veltman said.