Name: Madhuri Hegde
Title: Executive director, Emory Genetics Laboratory
Next-generation sequencing is not a standalone test, according to Madhuri Hegde. Instead, clinical genetics laboratories should use a combination of deletion/duplication arrays, targeted sequencing panels, and whole-exome sequencing to uncover the mutations that may be causing a patient's disease or disorder.
As executive director of Emory Genetics Laboratory in Atlanta, Hegde has been an active user of both technologies. In partnership with Oxford Gene Technology, she has designed nearly two dozen molecular arrays that focus on particular diseases or syndromes, beginning with a chip for Duchenne muscular dystrophy that OGT launched as a catalog product in 2009 (BAN 11/3/2009).
During an OGT workshop at the American Society of Human Genetics annual meeting in Boston last week, Hegde gave an overview of EGL's activities and its combined use of microarrays and sequencing to offer what she believes is comprehensive testing. BioArray News sat down with Hegde after her presentation. Below is an edited transcript of that interview.
As a clinical genetics laboratory, how do you differentiate yourself from other labs?
EGL focuses a lot on comprehensive testing. The idea is that we don't want to offer assays that are not comprehensive and there is no point in that because you are missing mutations. So we have separated ourselves in several ways. We are doing comprehensive assays, including arrays to cover mutation spectrum of the genes included in the panel. We have 19 focused panels. We also offer whole-exome sequencing. We are the first clinical lab to open up our database to the public, through an online tool called EmVClass. We will also upload this data to ClinVar. Everything I showed during my [ASHG] talk is all in the database and is free for download with no registration required. All data, from arrays and sequencing, gets integrated into our internal database, EmBase, first, and then displayed through EmVClass.
During OGT CEO Mike Evans' introduction, he said several times that arrays and sequencing are complementary. How do you feel about that statement? Do you think there are people who don't see them as complementary?
There is a lot of hype around next-generation sequencing. Next-gen sequencing is not the only solution as it does not detect all types of mutations. There are many complementary assays that go with it. And I think that is what is missing in the presentations about next gen and the overall detection rate. We are becoming so 'one track' now – next-gen, next-gen. What you can do clinically and what you can do on a research basis are two different things. And clinically you are also bound by turnaround time. And that is why I presented that you can detect copy number variation by next-gen, but you need to confirm it in the patient. The need for confirmation adds time and additional expense. We have validated deletion/duplication assays on microarrays, which saves time.
You were using molecular arrays before you got into next-generation sequencing. Did you immediately see the fit between the two technologies?
When I came from Baylor to Emory, the first assay I developed was for DMD. Seventy percent of events are deletions and duplications. So, the obvious assay to develop was a standard array for clinical use. In DMD, it is also important to determine the exact number of exons deleted or duplicated and how close you are to the break point. Some events can be complex. To give you an example, there was a kid who got enrolled in a clinical trial with a certain, documented deletion, but when we ran the array [it] showed that the break point was so close to the flanking exon, it was actually deleting that exon, too, which we confirmed later by cDNA sequencing. But it was not detected with PCR or [a multiplex ligation-dependent probe amplification] assay because the exon could not? be amplified. And it changed the deletion event from being out of frame to in frame. We contract with pharmaceutical companies, and we are getting into DMD clinical trials to use the array as the technology to capture the precise event. So we started putting more genes on the DMD array. Because my research is in neuromuscular disorders we did those first, and then expanded into X-linked disorders and autism because others in our department are interested in autism. And now there are about 22 molecular arrays.
And all of these have been made available via Oxford Gene Technology. Why did you decide to work with them?
We did a side by side comparison of all three platforms – NimbleGen, Agilent, and OGT. We presented a poster at ASHG about three or four years ago, and that is when we decided to switch to OGT. The number one thing was that we saw the improvement of the definition of the deletion and duplication by the design strategy and another thing was the flexibility, because I keep on adding genes, and we maintain a dialogue, which helps. We tried other platforms, but this just worked well for us.
Improving the platform's performance has been a very mutual exercise. It has helped in defining these arrays and their clinical utility. Because adding the array pipeline to the NGS pipeline to do integrated testing is a big decision. And it has really helped us a lot. We are also focusing on integrating reporting so that we do not do a sequencing report and then a follow up deletion/duplication report. It is one single report with integrated data.
But what does comprehensive testing really mean? Why do you feel certain or secure that you are giving a patient everything with these two technologies?
That's a great question. At EGL, the labs are upstairs and the clinics are downstairs. There is a lot of interaction. Clinicians are talking to us about the clinical presentation of a patient. They want to be sure before they go on to the next diagnosis that they have looked at everything. I can give you plenty of examples where samples get sent to us where only sequencing was done. When a clinical test is ordered, the physician should be asking, "Does your lab offer comprehensive testing?" Because if I report a negative result, then they can think of panels or exome sequencing. And that is what I think is missing. For recessive disorders when a single mutation is found, an array can be done to detect deletions/duplications. But in many cases it does not get done so we offer it as a comprehensive test so that testing is complete.
So do you think there a tendency out there among some labs that more sequencing is the answer?
That's how it is advertised, essentially. I think there is an education aspect here. The questions to ask are, "How comprehensive is it? How much am I going to miss if I don't do this?" For example, we complement X-linked intellectual disability next gen panel with del/dup array and assays for trinucleotide repeat disorders, Fragile X syndrome, [and] FRAXE syndrome. Another example is the short stature next gen test, which is complemented with deletion/duplication array and methylation assays. Some syndromes cannot be detected by next-generation sequencing. You have to do a methylation assay. So if you have a kid with one of these syndromes, you should do the obvious test first. As the head of a lab that offers a clinical test, it is my job to ensure that the technology I am using is covering the mutation spectrum.
You also provided an update on the MedicalExome project during your talk. How is that going?
We are developing a medical exome in collaboration with Avni Santani at [the Children's Hospital of Philadelphia] and Birgit Funke at Harvard [Laboratory for Molecular Medicine]. The idea is to offer complete coverage of all medically relevant genes, about 4,600, from the 22,000 genes covered by exome sequencing. We are still sequencing the exome but in this project we are focusing on comprehensive gene curation efforts in collaboration with the ClinGen project and also technical improvement of the assay to offer complete coverage for the medically relevant genes. It is possible that we will have to use complementary methods to detect those changes not currently detected by NGS.