Skip to main content
Premium Trial:

Request an Annual Quote

Somalogic Using DeCode Deal to Drive Development of Clinical Proteomics Pipeline


NEW YORK (GenomeWeb) – Somalogic plans to use its recently announced collaboration with DeCode Genetics to drive development of its clinical proteomics pipeline.

Under the agreement announced last week, the companies are using Somalogic's SomaScan proteomics platform to characterize 40,000 of DeCode's patient samples, with DeCode using the resulting data for its drug discovery and development work and Somalogic using the data to develop clinical applications for SomaScan.

The deal is part of Boulder, Colorado-based Somalogic's recent shift from offering SomaScan on a fee-for-service basis to a collaboration-based model in which it retains access to the data generated by its work with outside parties. That collaboration-based model is key to accessing the large and diverse patient cohorts the company needs to support its goal of commercializing SomaScan as a clinical platform, said Stephen Williams, Somalogic's chief medical officer.

The SomaScan platform uses the company's aptamer-based affinity reagents called Somamers to measure proteins in patient samples, typically blood. Somalogic launched SomaScan in 2012 and since then has derived the bulk of its revenues from service work using the platform, with the pharma industry being one of the firm's largest customers.

In the last year, though, the company ended its fee-for-service business to focus instead on projects that it believes will push forward its clinical goals.

Previously, "We effectively sold [SomaScan] data for money, but we didn't get to see the data ourselves," Williams said. "And what we realized was that despite bringing in revenue, this wasn't actually advancing the key aims of the company."

"We [now] give our collaborators the protein data and they give us the clinical data," he said, "so we can both learn and we both get permissions to use those data for our relative objectives."

He added that to entice potential collaborators, Somalogic has made its full panel of 5,000 protein measurements available at roughly half its previous cost. Under its fee-for-service business, the company typically offered a smaller 1,300-protein SomaScan panel. It runs the 5,000-plex panel out of its CLIA facility.

With regard to the DeCode agreement, Somalogic is getting access to what Williams said is "roughly 800,000 participant years of follow-up data," that it will use to "develop or validate protein models that will help predict future health outcomes."

"These collaborations are really from our perspective subsidized knowledge generation," he said. "We're not making a big profit margin. The whole point is to expand our knowledge base so that we can scale the [clinical] product development in a way that dwarfs what other people have been able to do, and it's more than we would be able to do if we self-finance all of our internal discoveries."

The company's need for access to large patient cohorts stems in part from its approach to clinical test development, which differs from that traditionally pursued by clinical proteomic and protein biomarker firms.

Instead of identifying a fairly narrow intended use and building and validating tests for that purpose, Somalogic hopes to use the depth and throughput of its SomaScan panel to develop and sell what Williams called "health insights," multiple protein measurements that provide information on a variety of aspects of patient health or disease status.

"We want to deliver all of the relevant information that a person or a care team might want to know to manage a patient's health situation," Williams said.

And while traditional proteomic test development involves collecting carefully curated case and control samples to discover and validate tests addressing narrow clinical questions, Somalogic has "gone the opposite direction," Williams said, developing its tests in larger, more heterogeneous populations "with all the co-morbidities and different drug prescriptions represented in the same way they will be in the population that [ultimately] receives the product."

Such an approach sacrifices statistical power, he noted, which is why the company is pursuing access to large patient cohorts through its collaborations.

Somalogic is piloting this approach through a collaboration with the Leeds Centre for Personalized Medicine and Health in the UK, where it is looking to see whether use of SomaScan data can improve compliance with a diabetes prevention program.

Williams said the company is currently delivering eight "health insights" on these patients, using the SomaScan platform to assess questions including risk of a cardiovascular event, risk of developing diabetes, kidney function status, liver fat status, body composition, smoking behavior, and alcohol consumption.

How the company will package and commercialize such data remains an open question, Williams said.

"Do we really have a single product, which is a holistic health assessment, and you let the patients and the doctors choose what goes in it?" he said. "Or are we rigidly fixing bundles of these models together so that, if you're a diabetic, you get this bundle, if you have coronary heart disease, you get that bundle?"

"We haven't [determined] that yet," he said. The company plans to launch an initial clinical product based on the SomaScan platform next year.

Somalogic is also still determining what the exact selling points to health systems and patients will be. In the case of the Leeds project, Williams noted that the study is not designed such that the company will be able to say its tests result in fewer diagnoses of diabetes down the road but rather is intended to show that the SomaScan data leads to greater patient engagement and compliance.

"There are different options for evidence of impact," he said. "Is it about the patient experience? Is it about the provider experience, that somehow you've saved them a lot of inconvenience and time that they would otherwise have had to spend? Or is it about health economics? Or is it about outcomes?"

"All those are different kinds of evidence of impact," he added. "We're really still working out the details of the nature of the evidence and the weight of evidence we would need to generate in order to get paid."

While Williams said that some SomaScan-based tests including those for predicting cardiovascular event risk and diabetes risk being used in the Leeds collaboration outperform existing available measures, the company has published little peer-reviewed data to date demonstrating the clinical performance of its SomaScan tests.

In 2016, the company published a study in the Journal of the American Medical Association detailing the performance of a SomaScan-based test for predicting the four-year probability of patients with coronary heart disease of experiencing myocardial infarction, stroke, heart failure, or death, finding that the test provided a slight improvement over the commonly used Framingham secondary event risk score. 

A version of that test is currently in clinical trials led by Nelson Trujillo, a cardiologist at Boulder Community Health, and was featured in a recent New York Times Magazine article by Michael Behar. In the article, Behar recounts his mother's unexpected death from a heart attack and his decision to seek out the Somalogic test.

Behar's initial test results gave him an 11 percent chance of experiencing a cardiac event within five years. A year later, after having lost eight pounds, gone on statins, "mostly eliminated red meat" from his diet, "added high-intensity interval training" to his workouts, and undertaken a five-day fast recommended by Trujillo to reduce inflammation, Behar's test score was unchanged.

While this might call into question how sensitive the test is to changes in patient cardiovascular health, Williams suggested that the one-year interval was too short a time to have observed the effects of the interventions Behar took on his risk score.

"I'm not so worried that his risk increment was not detectable, because he probably didn't wait long enough," he said.