Skip to main content
Premium Trial:

Request an Annual Quote

Fred Hutch Spinout Ozette Bringing AI to Single-Cell Omics Data Analysis


NEW YORK ─ With $6 million in new seed funding, Ozette, a spinout from the Fred Hutchinson Cancer Research Center and the Allen Institute for Artificial Intelligence (AI2), is looking to tackle the immune monitoring and immunotherapy market.

The Seattle-based company's Immune Monitoring Platform uses artificial intelligence to speed single-cell omics data analysis and provide a high-resolution view of subjects' immune systems.

Ozette plans to use its seed funding round, which was led by Madrona Venture Group with participation from investors including AI2 and Vulcan Capital, to further develop its platform.

The company is currently focused on single-cell proteomic data, primarily in the context of cancer immunotherapy, said Greg Finak, cofounder and CTO at Ozette and a former senior staff scientist at Fred Hutchinson. He said that much of his single-cell work previously had focused on infectious disease.

"That's basically what the past 15 years of research that myself and the team that has come over from [Fred Hutchinson] has focused on," he said, adding that among the group's efforts was using single-cell proteomics data collected primarily via flow cytometry to identify markers of outcome and response to HIV vaccines.

Finak said that a focus on proteomics made sense for the company given that proteins are the primary functional units of cells and the target of most drugs.

Ozette aims to apply the tools developed through this work primarily to drug development and identification of markers of response in cancer immunotherapy. According to Ali Ansary, the company's cofounder and CEO, the firm is currently working with around half a dozen of the major pharma companies with immuno-oncology programs, though he declined to name any of these collaborators.

He noted that while cancer immunotherapy is the company's initial area of focus, it may in the future expand into other areas included infectious disease.

Finak said Ozette is working with pharma and other partners to build out its analysis platform using single-cell data from these collaborators and other sources. He said the company plans in the near future to begin generating its own single-cell data in house. Fundamentally, Ozette's pitch is that its platform will allow customers to make better use of the vast amounts of single-cell data that researchers have generated over the decades using tools like flow cytometry and that is now rapidly proliferating with the adoption of new technologies like mass cytometry.

"What makes the single-cell proteomic data that we are focusing on now sort of unique is that the [flow cytometry] technology has been around for 50 years," Finak said. "There is an immense amount of data that is already out there."

Tools for analyzing this data have been one of the major bottlenecks within the field, he said. "We think we can capitalize by starting with a retrospective look at a lot of the data that has been published. We know there are a lot of findings and information that has basically been left on the table."

He said that Ozette has identified a number of potential biomarkers by retrospective analysis of old datasets.

"We think there are discoveries to be made by using this technology on data that is already available, and that will help inform where we want to focus as we generate our own data," Finak said.

He noted that the single-cell proteomics has advanced almost exponentially over the last decade as vendors like Fluidigm have developed new technologies like the CyTOF capable of analyzing dozens of proteins per cell. As such, he noted, the number of

possible cell types that that can be identified has grown.

"And now we really have the ability to start to see these rare cell types in the immune system that actually give us answers to questions like why a drug works and why it doesn't," Finak said.

At the same time, this has further heightened data complexity and data-analysis challenges, Finak said. "The reality is that with single-cell datasets, as popular as they have become, scientists are only extracting about 10 percent of the information, because even though they are measuring so much they are only able to resolve a small fraction of the cell types in [their data] in a way that is reproducible and reliable," he noted.

Ansary said the company has had some early success in analyzing datasets from skin cancer clinical trials, identifying a set of memory T cells expressing a particular combination of proteins that appear predictive of response to therapy in both retrospective and prospective validation work.

Ozette has also partnered with several cancer centers to collect single-cell and clinical data from patients undergoing treatment, allowing it to use its platform to look for markers of response in these subjects.

Finak said that in addition to flow and mass cytometry data, the company is also beginning to look into using transcriptomic data as well as mass spectrometry-based single-cell proteomic data, which still suffer from low throughput.

Ansary said the company is also interested in working with spatial omics data, though he said this was further off in the future. Single-cell spatial omics has grown quickly with vendors like Fluidigm, IonPath, Akoya Biosciences, NanoString, and 10X Genomics marketing or developing spatial omics technologies. Additionally, mass spectrometry imaging continues to advance, providing another source of spatial omics data.

"The goal is to be able to have a truly multi-omic platform," Ansary said. "We are waiting for the science and the instrumentation to catch up, but we are seeing every year manufacturers and companies continue to push instrumentation further and further."