Skip to main content
Premium Trial:

Request an Annual Quote

Bioinformatics Startup Looks to Break Data Silos for Precision Medicine


CHICAGO (GenomeWeb) – A startup run by the former Intel executive who helped build two major cancer informatics systems has big plans to support precision medicine by helping clinicians and researchers break down data silos on multiple fronts.

Omics Data Automation, founded in June 2016, has one commercial product out, another one set to be released this quarter, and an open-source community in development. The Beaverton, Oregon-based company is doing all this with barely $1 million in funding, nearly all of it from grants.

CEO and Cofounder Ganapati Srinivasa started Omics Data Automation following a long career at Intel, most recently as head of the microchip giant's Xeon server business. "I architected all the multicore processors on the server line," Srinivasa noted.

As part of that job, Srinivasa was the primary architect of the Intel-OHSU Collaborative Cancer Cloud, a platform for hospitals and research institutions who are interested in sharing private genomic, imaging, and clinical datasets in a secure environment, developed in conjunction with researchers at Oregon Health & Science University's Knight Cancer Institute. Other partners in the Collaborative Cancer Cloud included Dana-Farber Cancer Institute and the Ontario Institute for Cancer Research.

Srinivasa also was a central figure in Intel's collaboration with the Broad Institute to develop GenomicsDB, an open-source method for storing and processing variant data. Hardware and software optimizations Intel introduced prior to GenomicsDB also helped the Broad get its Genome Analysis Toolkit-based whole-genome sequencing time down to just 24 hours using a single Xeon server in 2014, Srinivasa said.

While building the Collaborative Cancer Cloud, Srinivasa got to work closely with Christopher Corless, who runs OHSU's Knight Diagnostic Laboratories, and with Kemal Sonmez, now the computational lead of the Cancer Early Detection Advanced Research (CEDAR) Center at OHSU's Knight Cancer Institute.

Srinivasa took a sabbatical from Intel in 2016 and in his words, "soul searched." After 24 years with the company, including 15 in charge of server architecture, he decided to try something "more fun," namely bring biology and computer science together in the emerging field of precision medicine. One week after leaving Intel, he started Omics Data Automation with Corless and Sonmez. In addition to his OHSU job, Corless also serves part-time as ODA's chief medical officer.

The lead product is known as the ODA Framework, which breaks down silos by creating a "data lake" to pull in relevant information from genomic, imaging, and clinical sources, including electronic health records.

Countless bioinformatics vendors have promised to tear down data silos. Srinivasa said that ODA tries to differentiate itself.

"You can always have three databases and basically query from one database to another database," he said. "We have gone one step further and created a unified data structure and model that can create the image of the patient as a single entity."

Plus, ODA is able to scale its architecture to handle cohorts of patients, he added.

"As architects and engineers, we started from the ground up basically using optimized databases for both imaging and custom databases, not something that's already out there. So these are concentrated databases specifically for that data type which are optimized for that," explained ODA Director of Business Development Aleksandr Shargorodskiy, who previously was an Intel hardware engineer.

"Then we created a framework on top of it which ingests all that data so you can then compute parallel across all data types across the genome, actually eliminating the barriers, not just querying from silo to silo," Shargorodskiy said.

"All the data types are recognized appropriately by the system so the semantic syntax is clear," Srinivasa added.

Development of the framework was funded by a one-year, $224,903 National Science Foundation Small Business Innovation Research Phase I grant awarded in mid-2017. With that grant, ODA was able to demonstrate the integration of genomics and imaging data.

A year later, the company received a two-year Phase II grant worth $747,352. With that money, ODA is supposed to demonstrate its technology at scale with analytics capabilities and additional data sources, including EHRs and histopathology images.

From the beginning, Srinivasa made the decision to seek grant funding rather than seek out venture capital. The founders put in some money and the company also received two small grants from local funding sources in Oregon. All told, ODA has raised just about $1 million to date.

Another strategic move the company made was to support GenomicsDB as an open-source product. "We were working on products, but we were also working on a parallel strategy that we need to put out some open-source software to get accepted in the community," Srinivasa said. "GenomicsDB was our serious work for the entire community."

ODA actually has detailed some of its own people to GenomicsDB development and maintenance of the website.

In the Phase I grant, Omics Data Automation customized its framework for NSF. The fruit of that labor should hit the commercial market this quarter, according to Srinivasa, following testing by two pilot customers, UCLA and the Children's Cancer Therapy Development Institute, a nonprofit in Portland, Oregon.

"Our current generation of products is data aggregation, automation, analysis, tumor board reporting, and cohort finding," Srinivasa said. "We will be launching our products, which are basically our framework and packaged product for tumor boards and cohort analysis in the hospitals and clinics," Srinivasa said.

Also before the end of Q1, the company expects to launch more GenomicsDB use cases.

Even as the company gets into its Phase II SBIR grant, it also is thinking ahead. "The second generation of products will be addressing tumor mutation burden and tumor microenvironment, so [hospitals] can create protocols and have some quantitative factors associated with the protocol for the right type of treatments for the clinicians," Srinivasa said.

The company's next clinically oriented product is further down the road, he added.

In the near term, ODA is taking its technology overseas, working with an unspecified "major institution" in India that, according to Srinivasa, wants to create an Indian version of the Cancer Genome Atlas. Expect details on that deal by the end of January, he said.