Illumina has explored the clinical interpretation of results from its Individual Genome Sequencing service in a small group of professionals and has added clinical reporting as an option for its service, Clinical Sequencing News has learned.
Illumina has been offering its personal whole-genome sequencing service since 2009 (IS 6/16/2009). The test needs to be ordered by a physician and the sequencing is performed in Illumina's CLIA-certified and CAP-accredited Clinical Service Laboratory. Until recently, the company only delivered a list of annotated single nucleotide variants and left the clinical interpretation to third parties and informed doctors.
But, according to Tina Hambuch, Illumina's scientific liaison for its clinical laboratory, the company decided that it also wanted to "tell patients and physicians how the data answers their questions."
Earlier this month at the Personal Genomes and Medical Genomics conference at Cold Spring Harbor Laboratory, Hambuch presented some results from its pilot project.
Participating in the project were 52 professionals who are active in the field of human genetics and genomics and were presumed to be healthy. Illumina sequenced their genome -- or the genome of a patient they chose to submit instead of their own genome -- in its CLIA laboratory. Last month, participants gathered at an Illumina-sponsored symposium, called "Understand Your Genome," to discuss their results (CSN 6/27/2012).
On average, Illumina researchers detected 3.3 million variants in each individual, about 21,000 of them in coding regions, and about 9,600 representing non-synonymous changes.
Partly because that number was too large to interpret in each person, they decided to focus on 344 genes associated with monogenic diseases. These included genes involved in adult-onset conditions with moderate to high penetrance as well as genes that are part of carrier screens. They were chosen based on guidelines from professional organizations and state recommendations for carrier status testing. "These are probably the most well understood genes out there," Hambuch said.
Another criterion was to only include those genes that would be completely covered by Illumina's sequencing and annotation process, which currently only analyzes single-nucleotide variants. This eliminated, for example, the fragile X mental retardation gene, because the disease involves trinucleotide repeats, as well as monogenic diseases usually caused by duplication events.
On average, the Illumina team discovered about 1,300 single nucleotide variants per individual in those 344 genes, although by the time they had reached the last genome, they only found about 50 variants they had not seen before.
About 30 to 40 variants per person were categorized in the Human Gene Mutation Database as deleterious mutations and were manually assessed by the researchers, who reviewed the underlying literature, searched locus-specific and allele frequency databases, and used functional prediction algorithms. A team of geneticists and genetic counselors then evaluated all the evidence.
The analysis required about 19 hours of manual review per genome, which Hambuch said is expected to decrease over time as analysis tools get better and their experience grows, aided by other laboratories willing to share their experience. One reason for the lengthy analysis is the fact that the genes they chose are well characterized and have a lot of literature associated with them.
And although there are many software tools to aid with steps in the interpretation process, the final assessment of weighing the evidence will be difficult to automate. "There really just isn't a software that can do that at this point in time," Hambuch said. "And in fact, that's not even easy for people to do. We actually did have this team of people, and we would really argue about these things. Some of these are not easy to decide."
The team found that on average, individuals were carriers for one to two diseases. At least five individuals were not carriers for any disease associated with the 344 genes analyzed, while a few were carriers for up to five diseases.
About every tenth participant had a mutation in a gene associated with an adult-onset condition that "looked really pathogenic," Hambuch said, though in many cases, those variants were associated with mild pathogenicity. In one case, they found a novel stop mutation in the BRCA gene, and it is now up to the participant to follow up on this result.
Konrad Karczewski, a graduate student in Mike Snyder's lab at Stanford University, participated in the pilot project and had his own genome sequenced. "Their analysis was fairly conservative, and understandably so, as they followed strict guidelines for variant assessment," he told CSN via e-mail. "For my genome (as for most participants), their analysis found a couple of such carrier variants, but nothing with any family history. Their analysis can be very useful for individuals with rare diseases and clinicians analyzing data for patients, but it is not particularly interesting for healthy individuals," he said.
But Karczewski, who recently co-authored a guidebook entitled "Exploring Personal Genomics," plans to do his own analysis. "Personally, I'm most excited about the raw data, which I've started analyzing with open-source tools, including ones we are developing [at Stanford], such as the Interpretome platform, which we are expanding to whole genome data."
While this first round of analysis did not include any medical history from participants, in a second round, Illumina is now analyzing their data with regard to family history of disease or to solve a specific clinical question. "Now we're saying, given this clinical background, can we find anything that would make sense for what's going on in these patients?" Hambuch said.
In addition, Illumina plans to analyze the genomes of another batch of 50 individuals next year and to hold second symposium in May. The company also aims to expand the number of genes interpreted, and to re-analyze the data of the first 52 participants for those additional genes.
It is also working on extending its interpretation to variants other than substitutions, which Hambuch called one of her "highest priorities." For example, the company plans to incorporate information from microarrays that it already runs on each individual sequenced for quality control in the downstream interpretation analysis. In parallel, it is exploring algorithms that can call insertions and deletions confidently. "What we always have to balance is trying to provide as much information as possible but ensure that the quality remains high," Hambuch said.
Illumina is now offering the interpretation of the 344 genes as an option for its Individual Genome Sequencing service, and is able to customize that gene set somewhat on a case-by-case basis.
"The main point of this [pilot study] was just to try to understand what the community needs, and the scope of the problem of trying to give people genomic-level information," Hambuch said, as well as to showcase Illumina's current process in order to obtain feedback.
"It's got to be a community effort. The difference between a genetic test and a genomic test is scale, and that is kind of true for the number of people who need to be involved, too," Hambuch said.