Personalis has launched an early-access program for its "Accuracy and Content Enhanced," or ACE, Clinical Exome test to diagnose rare Mendelian disorders, which it hopes will increase the diagnostic yield of conventional diagnostic exome sequencing.
Last week at the American Society of Human Genetics annual meeting in Boston, the company presented first results from three pilot collaborations with a group at the University of California, Los Angeles, and two groups at the University of California, San Francisco, that used its enhanced exome service to solve more than half of 30 cases studied so far.
By comparison, most other groups providing clinical exome sequencing have shown diagnostic yields on the order of 30 to 40 percent. However, the number of cases analyzed by Personalis is still small, and it is unclear if the increased diagnostic rate will hold up in the future.
"If we can continue this, and we can show on a larger scale that we can get to 50 percent diagnostic yield or even higher … then I think it will be quite a validation that what we're doing is actually better," John West, the company's CEO, told Clinical Sequencing News.
The Personalis ACE Clinical Exome combines an exome sequencing protocol that fills in gaps of coverage for more than 7,000 important genes with sequencing-based genome-wide structural variant detection.
"This is really the first clinical exome product that also has the ability to look at genome-wide structural variation," West said, noting that this normally requires a separate test, such as array-CGH.
The ACE Clinical Exome is less expensive than separate exome and array-CGH tests, he said, and saves ordering doctors time because they only need to look at the results of one test.
Also, he expects Personalis' structural variant assay to be better at detecting small changes than array-CGH, although his team has not yet performed a detailed comparison with array-CGH, which it plans to do as part of the early-access program.
During the early-access phase, the test will have a turnaround time of eight to 12 weeks, from receipt of a sample to issuing a clinical report. Most of that time will be taken up by the analysis and interpretation. Customers will receive "time slots" for their samples, enabling them to track their samples.
The price for the test is currently $7,500 per case, which will mostly involve sequencing parent-child trios. "If the parents are available, we would like to sequence them, so we can show what fraction of the time it did make a difference to have the parents," West said.
The company currently uses Illumina HiSeq 2500 and MiSeq platforms to generate the exome data and to fill gaps in GC-rich regions after doing a targeted capture with a sample prep method optimized for high GC content.
At the moment, more than 5,000 genes that have been linked to medical conditions have more than 99 percent of their bases covered with at least 25-fold sequence coverage.
For structural variations, Personalis uses a genome-wide sequencing-based method that has not yet described in detail, West said.
To analyze the data, it employs an optimized pipeline for alignment and variant calling that incorporates an in-house enhanced reference genome, and an annotation pipeline that uses content from proprietary disease variant and pharmacogenomics databases as well as public and commercial databases (CSN 3/6/2013).
The clinical interpretation is primarily driven by phenotype. To this end, the firm has built its own database that links clinical features with genes shown in the literature to be associated with specific phenotypes, West said.
Based on the patient's clinical phenotype, Personalis queries several databases to generate a comprehensive list of candidate genes that are ranked based on how well they fit the phenotype. The company also ranks genotypes by their ability to explain the family disease history.
The company reports secondary findings in accordance with the guidelines from the American College of Medical Genetics and Genomics but gives patients an option to opt out of these findings.
All results are validated by capillary sequencing or, in the case of some structural variants, by real-time PCR, before they are reported to ordering physicians or clinical researchers.
All assays run in Personalis' laboratory, which is currently CLIA-certified in all states except Florida and New York and is working on becoming CAP-certified.
The company plans to launch the clinical exome more broadly in the first quarter of 2014, after scaling up its technology and automating aspects of the process that are still manual at the moment, West said.
For the early-access program, it is looking for groups that have "serious clinical cases" and are interested in publishing the results. It plans to take on a "modest number of customers" with a total of several hundred cases.
Among the goals of the program is to characterize the diagnostic yield of the assay, and whether it performs better for some medical conditions than others, West said. The company will also try to tease out which of the components that differentiate it from other clinical exome providers – be it the enhanced exome, enhanced reference, or one of its proprietary databases – make a difference for the results.
The three pilot collaborations with UCSF and UCLA involved a total of 30 cases to date, including individuals and families with hereditary cancer syndromes, inherited eye diseases, and multiple congenital abnormalities.
For 16, or more than half the cases, Personalis was able to report a finding. These included 12 genes or genome regions that had previously been described in patients with similar phenotypes, and two novel genomes for which strong evidence exists that they are linked to the disease in question.
While more than half of the causative mutations were missense mutations, the company was also able to detect an exon deletion and a large-scale deletion.
Almost half of all cases solved by Personalis were family trios, and about a third were families with multiple affected individuals.
For the UCLA group, Personalis was able to solve nine out of 10 cases of patients with hereditary eye disease. "We had really hand-picked cases, which were very likely to have a good result because we've had them for many years," said Samuel Strom, an assistant professor in the department of pathology and laboratory medicine at UCLA. He added that the success rate is likely to be lower in the future.
For one family with three brothers suffering from early-onset cataracts, he said, the company found a homozygous mutation in a gene that had been described in the literature once, for a case with a very similar phenotype. However, no genetic test for that gene is currently available.