NEW YORK (GenomeWeb) – Demonstrating how large-scale sequencing projects can be translated into clinical applications, a group of researchers in the UK has used sequencing data from 5,000 whole genomes, half of them from patients with blood or immune diseases, to develop a targeted NGS panel for heritable bleeding, thrombotic, and platelet disorders.
The assay, developed by researchers within the ThromboGenomics consortium, is being offered for free to clinicians within the UK's National Health System throughout 2016, and is run out of a lab at the University of Cambridge.
The sequencing work and subsequent assay design were part of a UK National Institute for Health Research BioResource rare disease project that aims to sequence the genomes of rare disease cohorts to identify the causative genes and use the findings to develop better diagnostic tests. Within that project, which has the capacity to sequence 10,000 genomes, there are 12 subprojects under an umbrella organization known as Bridge, including the bleeding, thrombotic, and platelet disorders.
This fall, the data generated from those 12 Bridge subprojects will be merged with the UK's 100,000 Genomes Project data through an opt-out process.
The consortium described the validation of the test in a recent study published in Blood. Willem Ouwehand, head of the NHS blood and transplant research group and chair of the ThromboGenomics consortium, told GenomeWeb that the group was able to go from sequencing to a clinical test in a relatively short period of time due to the amount of data sharing between laboratories and the large numbers of sequenced genomes. "We've sequenced thousands of patients and then shared the data freely, between the laboratories," he told GenomeWeb. "So it's much quicker to get evidence to say that this gene is relevant." Once the group decides on a gene, he said, it takes about six weeks to get it on the platform and validate it.
In 2009, just before NGS technology really took off, Ouwehand said, there were about 34 genes known to be responsible for these inherited bleeding, thrombotic, and platelet disorders. Now, there are 78 genes included on the assay. And, said Ouwehand, many of those genes have been added since the initial validation of the test, which included 63 genes. He anticipates that as many as 200 genes could be included in the next two years.
The Bridge study's sequencing efforts discovered about 15 to 20 genes, Ouwehand said, and the remainding ones were added to the platform based on other recent discoveries that the group came across through literature searches.
In the validation study, the group ran the assay on 300 individuals — 159 with known pathogenic variants, 61 with a suspected but undiagnosed molecular etiology for their disorder, and 80 patients with no suspected disease etiology.
The researchers used Roche's NimbleGen capture technology and sequenced the DNA on Illumina's HiSeq 2500. The assay targets all exonic regions of the genes and some introns and untranslated regions. Average sequencing coverage was 1,178-fold.
The team noted a couple of issues with the assay. GC-rich regions of the genome are known to cause difficulties with sequencing technology, and the researchers had trouble with one gene in particularly, with an exon that had high GC content and contained 20 pathogenic variants. However, by changing the polymerase enzyme in the library prep, they were able to improve the coverage of those variants, and all other variants located in GC-rich regions. Another issue was calling variants found in one exon of the VWF gene, which is linked to von Wellebrand disease. That exon is "perfectly homologous" with part of a pseudogene, the researchers wrote, and they were unable to call variants in that exon confidently.
In the validation, the researchers first looked at the concordance of their assay with the 159 samples that had known pathogenic variants. The variants included 119 SNVs, 19 indels, and 7 large deletions in 37 genes, and the platform detected all of those variants.
Next, they tested the assay on 61 patients who were suspected to have one of the disorders included in the assay. Suspicious variants were identified for 56 of those cases, including 29 pathogenic and 28 likely pathogenic variants.
For the five cases the assay was unable to diagnose, the researchers said it was unclear whether it was due to a lack of sensitivity of the assay because the variant was in a regulatory region or another area not covered by the assay, or whether those cases did not in fact have an inherited disorder, but an acquired blood disorder.
In addition, the ThromboGenomics assay was able to find causative variants in five individuals who had previously tested negative in Sanger sequencing tests for their suspected condition.
For the third group of uncertain cases, the researchers were only able to make a diagnosis in 10.5 percent of patients, or 8 cases, "which underscores the need for further research into the molecular etiology of uncharacterized" blood platelet disorders.
The authors also noted that the assay cannot identify inversions, which cause 45 percent of severe hemophilia A cases. As such, they said, a PCR-based test should be performed to rule those individuals out before running the NGS test.
Ouwehand noted that the development of the test was a huge improvement over current methods for diagnosing blood platelet disorders but still had a long ways to go. For instance, he said, individuals who have very rare disorders and where the clinician has no idea what it is are not likely to find a molecular diagnosis with this assay, and are better suited for whole-genome sequencing. However, for individuals where the clinician at least has evidence toward a specific etiology, the assay has about a 90 percent chance of identifying the causative gene. In addition, he said, prior to this assay, molecular diagnostic tests were only available for seven of the genes on the panel, and those tests were "prohibitively expensive," running upwards of $1,600 per gene.
The assay also has the ability to change patient management, Ouwehand said. For instance, for hemophilia patients, the sequence of Factor VIII and Factor IX genes predicts whether they will have an adverse reaction to treatment. Treatment for hemophilia patients who have a bad reaction to the drugs can end up costing $2 million per year, he said, so knowing how they will respond to treatment can both prevent negative side effects and help keep healthcare costs down.
In addition, a number of genes in the assay can have variants indicating an increased risk for leukemia if the patient's platelet count is low. Sometimes that risk is "so substantial that patients need preventative bone marrow transplants," Ouwehand said.
And finally, the sequencing assay can clarify the prognosis of patients with big platelets. Some individuals with big platelets will go on to develop kidney problems and early deafness, which can be predicted based on variants in the MYH9 gene, he said.
Ouwehand said that the group plans to keep developing the ThromboGenomics assay further. It continues to aggregate new discoveries and decides every three to six months whether there is enough evidence to add new genes to the assay.
Currently, he said, the assay is free for NHS clinicians, but starting in 2017, the laboratory would likely begin charging. Details on cost are still being worked out, he said, but the price will likely be around $200.