Genomic Health is beefing up its informatics portfolio as part of its efforts to use next-generation sequencing for biomarker discovery and diagnostic testing.
Genomic Health began using next-generation sequencing as part of its research program several years ago and plans to eventually move its entire molecular diagnostic platform from quantitative RT-PCR to next-gen sequencing. The company announced this week that it is "accelerating" its NGS program by moving it from research into clinical development.
Randy Scott, executive chairman of Genomic Health's board, told BioInform that the company is ramping up its informatics efforts in order to keep pace with its sequencing activities. The firm has embarked upon a multi-pronged informatics strategy that involves a combination of open source and in-house developed tools, as well as third-party software from two startups in which Genomic Health has taken a financial stake.
Scott said that Genomic Health is using existing open source algorithms where possible, as well as algorithms provided by sequencing vendors, such as Illumina's Casava for base calls and alignments.
The company also intends to build "customized" tools to make sense of genomic data in a way that helps clinicians treat patients. Part of its effort on this front includes developing improved methods for RNA-seq data normalization and other statistical parameters.
Genomic Health is also relying on its $2.8 million investment in two informatics startups that are developing software for the clinical diagnostics market. As reported by BioInform's sister publication Clinical Sequencing News, company officials disclosed the investments in an earnings call last May (CSN 5/4/2011).
One of the startups, Station X, is developing software called GenePool that is designed to support data analysis in large patient cohort studies, Scott explained.
The company will soon kick off a beta run of GenePool for pharmaceutical, biotechnology, and diagnostic companies, and is expecting to fully launch the software by the end of 2012, Scott said.
The other firm, Locus Development, offers clinical tools and services to enable personalized, predictive, preventative, and cost-effective healthcare.
Scott explained that while Genomic Health could have opted to develop all of its software internally "we have chosen to focus more of our proprietary effort in ... beginning to put the data into a clinical environment where biologists and clinicians and not just biostatisticians can actually analyze and work with the data."
The company is in the process of installing an in-house high-performance computing infrastructure that includes Isilon storage, Scott said, though it is also exploring cloud options so that it can accommodate collaborative research endeavors. The firm is currently using Amazon's cloud but doesn’t intend to restrict its activities to a single provider, he said
Genomic Health currently markets two tests — Oncotype DX Breast Cancer and Oncotype DX Colon Cancer — which are used to predict patients' recurrence risk in early-stage breast cancer and stage II colon cancer, respectively.
The breast cancer test also predicts the likelihood that patients will benefit from chemotherapy and helps physicians recommend treatments for individuals with certain kinds of breast cancer.
Last December, the company released the results of a clinical outcomes study it conducted to test its sequencing-based biomarker discovery approach. This pipeline relies on NGS-based whole transcriptome profiling of formalin-fixed paraffin embedded tumor samples and mutational analysis to identify cancer biomarkers in large clinical studies.
In the study, Genomic Health researchers found that the method was able to identify the same validated breast cancer genes used in its Oncotype DX breast cancer test.
The company’s researchers applied the approach to samples from a cohort of 136 breast cancer patients with tumor tissue that was obtained at the time of diagnosis. Prior to the study, the tumors had been analyzed by RT-PCR in the biomarker discovery phase of the development of Genomic Health's 21-gene Oncotype DX assay.
Samples were sequenced using an Illumina HiSeq 2000 instrument, which generated over 40 terabytes of data, Genomic Health said.
In addition to re-confirming the 21 Oncotype DX genes originally identified by RT-PCR, the study also revealed more than 1,800 coding and non-coding RNAs that are associated with breast cancer recurrence, Genomic Health said.
While RT-PCR allows the company to look at "a few hundred genes," next generation sequencing offers "a whole transcriptome profile," Scott said.
"Sequencing 40 million RNAs from each individual patient ... really is an incredible opportunity to do very exhaustive genomic analysis of tumors," he added. "We are really interested in applying that into the clinical world."
Genomic Health is now moving its NGS approach into formal clinical development and hopes to have a commercial test based on it within the next five years, he said.
"By the end of this year our goal is to have one of the most robust platforms in the world for doing combined whole transcriptome [profiling] and large panels of mutations simultaneously at low cost in clinical studies," along with an informatics infrastructure that can support large clinical and commercial programs, he said.
In addition to developing new genetic tests to guide treatment decisions, Genomic Health envisions building databases of genomic information that could be used to select individuals to participate in clinical trials or studies, Scott said.
'The Key to the Future is Bioinformatics'
Although software development isn't Genomic Health's core focus, Scott described bioinformatics tools and technologies as "critical" to the clinical diagnostic market.
In addition to bioinformatics, he said that firms who invest in NGS technologies also need to ensure they have the secure storage, network capabilities, and computing infrastructure needed to support the large datasets that sequencers generate.
Scott noted that once Genomic Health's NGS-based test moves into development, the company intends to conduct more clinical studies that will involve multiple collaborators.
"We have got to be able to not just generate the data for internal use but put it into an environment where we can get that data to the cloud on dedicated fibers and in a secure fashion, and then be able to share data with our clinical collaborators and partners," he said.
He also noted that the decreasing cost of sequencing is going to have a "phenomenal impact" on the diagnostic market. And although software for the diagnostic space is currently a "heavily misunderstood and under-invested field," that is likely to change as NGS continues its march into the clinic.
"The key to the future is bioinformatics," he said. "I think some of the stars over the next five years are going to be some of the bioinformatics and software companies building the infrastructure ... this is an exciting area for future growth."
Have topics you'd like to see covered in BioInform? Contact the editor at uthomas [at] genomeweb [.] com.