By Julia Karow
Amid a market that noisily trumpets every advance in human genome sequencing technology, informatics firm Omicia has been quietly working on ways to interpret genomic data by linking genetic variants to disease.
The company, founded in 2002, last month inked its first commercial partnership by joining the Genomic Cancer Care Alliance, which is spearheaded by Life Technologies. The initiative, for which Omicia will provide genome interpretation services for about 100 cancer genomes, aims to test whether whole-genome sequencing can improve treatment decisions for difficult-to-treat cancers (IS 6/8/2010).
Over the last six months or so, the 10-employee firm has matured its genome interpretation tools. "They are now ready, from an industrial point of view, to be applied," said CEO and co-founder Martin Reese. "We want to prove ourselves in these commercial relationships — right now, the [Genomic Cancer Care Alliance] is a first test case for us."
Omicia is entering the market as a growing number of human genome sequencing projects are getting underway that will require clinical interpretations. "This is an area of great interest among both companies and academics," said Russ Altman, a professor of bioengineering at Stanford University who was involved in interpreting the genome of Stanford professor Steve Quake, which was published in The Lancet earlier this year (IS 5/4/2010).
When Reese founded Omicia in 2002, he said he was hoping to translate results from whole-genome analyses into the clinic. At the time, he said, it was unclear whether the dominant technology would be genotyping, large-scale expression studies, or genome sequencing, but he knew it was going to require complex interpretations of groups of genes rather than a single gene for a given disease. Reese previously developed human genome annotation software at Neomorphic, a company he founded in 1996 that was sold to Affymetrix in 2000.
Initially, Omicia focused on mapping rare genetic mutations — such as those listed in the Online Mendelian Inheritance in Man database — to the human genome.
In 2008, the firm began applying its approach to whole human genomes, analyzing the genomes of Jim Watson, Craig Venter, and an African Yoruban HapMap sample that was sequenced on Life Technologies' SOLiD system for health-relevant variants listed in OMIM and the Pharmacogenomics Knowledge Base. In addition, it has been using its own curated database of human disease genes.
In 2009, Omicia contributed to Life Tech's publication of the African Yoruban HapMap genome, the first published human genome sequenced on the SOLiD platform.
Last year, Omicia, which has been largely funded through small business innovation research grants from the National Institutes of Health, received a $590,000 phase II SBIR grant from the National Human Genome Research Institute to develop a computational system to predict novel genetic disease associations, in collaboration with Mark Yandell's group at the University of Utah. Reese and Yandell know each other from Yandell's days at Celera, and worked together on genome annotations for the Drosophila and human genome.
Under the first phase of the NHGRI grant, the partners developed a "gene interference system" to predict candidate genes and mutations that may play a role in a disease. Yandell built a disease gene ontology, which Omicia will now test, initially in the area of cardiovascular disease. "The idea is to use the disease ontology and then come up with a very comprehensive cardiovascular profile," Reese said. According to the grant abstract, the researchers plan to conduct a study involving approximately 700 cases and 700 matched controls, testing their top variant candidates and comparing them to well-established genetic markers for the risk of myocardial infarction.
[ pagebreak ]
Under a $400,000 NHGRI stimulus GO grant awarded to Yandell's group and Omicia last year, the partners have also been developing software for the automatic annotation and targeted analysis of whole-genome sequence variation data. The software, called VAAST, for Variant Annotation, Analysis, and Selection Tool, will be "useable for data from all sequencing platforms" and is expected to "fill a huge void in the software landscape by helping individual scientists to extract meaningful results from whole genome variant files," according to the grant abstract.
Reese said the partners have already used the tool to analyze 10 published human genomes, a project they intend to publish later this year. The software, which Omicia will apply commercially, allows the company, for example, to select and analyze all genetic variants in an individual's genome that are relevant for a certain disease, such as breast cancer, Reese explained.
This year, Omicia, with the help of Illumina co-founder John Stuelpnagel, an advisor to its board, and Paul Billings, CSO of the Genomic Medicine Institute at El Camino Hospital, a board member, has focused on commercializing its technology, for which it sees growing demand as more human genomes are sequenced for clinical projects.
The partnership with Life Tech's Genomic Cancer Care Alliance — a two-year project — is "a very good model of how we are applying our technology to the commercial space," Reese said. "We are doing the technical interpretation of genomes," he said, and "the medical community is transferring that over into advice to the individual."
Over the next couple of years, the company plans to provide genome interpretation services for similar projects with a clinical focus. "After that, we have to see how the whole space and technology is developing. If it goes extremely fast, there might be other models that we should explore," Reese said.
Commercial partnerships, which give the company access to phenotyped samples, also serve to validate its technology. "It is a process for us to validate, and in that way, our knowledge databases gets better and better," he said.
Omicia is, of course, not the only company providing human genome interpretation services. Knome, for instance, has provided whole-genome sequencing and analysis services to individuals since late 2007, and more recently also to researchers. In addition, the company uses its KnomeXplorer genome browsing software to interpret sequence data for customers of Illumina's Individual Genome Sequencing service. Earlier this year, Knome also licensed its genome analysis platform exclusively to BioMérieux for use in the in vitro diagnostics market (IS 4/27/2010).
Reese declined to comment on how Omicia plans to differentiate its analysis service from Knome's, citing the fact that Knome has not published a description of its tools to date. Publishing in peer-reviewed journals, he said, is going to be "absolutely critical" for Omicia to establish its reputation within the academic community.
In addition to companies, academic groups have embarked on whole-genome clinical annotation projects, such as Stanford's analysis of Quake's genome. Reese said Omicia is currently analyzing the Quake genome using its own technology, "looking at how we would have done it, what the differences and strengths are."
The Stanford researchers said that clinically annotating Quake's genome took approximately 20 months of manual curation, but Altman told In Sequence this week that he and his colleagues are currently trying to automate the interpretations they presented in the paper.
Also, George Church's group at Harvard Medical School has created Trait-o-matic, a "community maintained, open-source web service that interprets personal genomes," according to the website describing the tool, which cross-references non-synonymous SNPs with multiple genotype/phenotype databases.