Celsius is translating single-cell genomic insights into precision therapeutics for autoimmune diseases and cancer.
CEO: Tariq Kassum, M.D.
Based: Cambridge, Massachusetts
Clinical focus: Using machine learning alongside single-cell sequencing to tease out genetic drivers of disease
The scoop: Because tissues are heterogeneous, containing many different kinds of cells, Celsius is pairing the buzzy machine-learning platform with single-cell sequencing. Like fruits in a smoothie, some types of cells may overpower others in a mixture. Looking cell by cell allows Celsius to scrutinize individual “fruit” to discover which genes determine what cells do and how they interact with each other, along with what goes wrong to cause disease.
What makes Celsius fierce: Celsius is using a new set of tools that it thinks will unlock some of the secrets of autoimmune diseases and cancer—notably cancers that do not have a druggable driver mutation, which make up around 70% of all cancers, CEO Tariq Kassum, M.D., says. And most current treatments, of course, target particular mutations.
Single-cell genomics and machine learning “go together,” Kassum says, as the former doesn’t work without a “heavy computational component.” In the last few years, the use of machine learning (ML) and artificial intelligence (AI) has dominated the biotech and Fierce 15 field, though some are still guilty of using them as buzzwords.
Terms like ML and AI “are easy to throw around," Kassum says. "But for us, it is central to what we do. Our first C-level hire was a chief data officer, a machine learning expert from Tesla.
"Ultimately, ML is a tool like any other," he adds. "To me it’s just an application of statistics, right at the far end of the scale that started out with me learning P-values at college, for instance, to the end where we have ML.”
So, what is Celsius doing with its machine learning? “You need data from machine learning to find those little needles in the haystack,” he explains.
Combining machine learning with single-cell genomics and sequencing—and "marrying that with an ace drug discovery team"—gives Celsius an opportunity to create therapeutics to hit targets that haven't been addressed before, he says.
“[A]s we’re doing single-cell genomics, which generates billion of data points, that quantity of data and that complexity does not lend itself well to normal modes of statistics," Kassum says. "So, if we need 99.9% precision, we need ML, because it’s far beyond what a human being can do.”
And on its other tech, Kassum explained: “Celsius has the ability to look at disease on a cell-by-cell level at an extremely high resolution [and] the ability to look at RNA expression. We are actually seeing the behavior of cells, rather than their somatic DNA programming. This makes single-cell RNA sequencing a uniquely beneficial tool if you’re trying to understand the driver of a disease in otherwise genetically normal tissue.”
The biotech’s platform does this by looking at patients’ tissue samples as the raw material. “We’re actually looking at these tissues to try and find out what’s driving disease on a cell-by-cell level,” Kassum said.
But often biotechs gain these tissues with little knowledge of their origins; Celsius has penned a series of pacts with academic and other institutions to gain a deeper knowledge of each patient's history through anonymized data, including what treatments they've had. This helps the biotech better understand the individual samples and how interventions are working within the context of that patient’s history.
Celsius got off a very healthy $65 million series A last year, and in July of this year it penned a new partnership with Janssen Biotech. Under that deal, Celsius is using its single-cell genomics and machine learning technology to identify biomarkers that predict patient response in a phase 2a study testing Janssen’s plaque psoriasis medication Tremfya and its immunosuppressive drug Simponi in patients with ulcerative colitis.
It’s a win-win situation for both, Kassum said, as Janssen can feed the “cellular insights” from the collaboration into its clinical work, and Celsius will use them to “seed” its platform, adding the data to its own database to interrogate in its own drug discovery efforts.
Celsius had its first big meeting under the Janssen collaboration at the end of August. “I have been extremely impressed by the collaborative spirit of that team,” Kassum said of Janssen, “and they are really excited about it as well, because we are helping Janssen figure out a new translational medicine approach,” which is a pretty big deal for an early-stage, private biotech.
At launch, Celsius had about 15 staffers and it has nearly tripled since then to 40. The company is looking to bolster its leadership, as well as its broader team; in particular, it is on the hunt for computational biologists. If all goes well, the company should be well into the 50-employee range by the end of 2019.
And the plan is to stay independent, while also working with biopharmas and seeking out more cash. “We’re a biotech company, so we have to be opportunistic; we spend money, that’s what we do, so we’re already talking about pharma collaborations,” Kassum said. It's important to avoid grabbing at whatever comes knocking and wait for the right partner: “We know how bad a dysfunctional collaboration can be,” he said, “and we all know what happens when it goes wrong.”
Another funding round is also in the cards, with a potential IPO a little further down the line, when it has a series of strong milestones to attract public investors. And the idea here isn’t to build up as much as possible just to sell off to the highest bidder.
“The plan is to become a fully integrated drug discovery and development shop. The long-term vision is to make this into something special and valuable; I don’t want us to just package it up for sale. We have a real shot at building something bold and unique, so we should try and do it from the ground up.”
Investors: $65 million series A round in 2018 led by Third Rock Ventures with participation from GV (formerly Google Ventures), Heritage Provider Network, Casdin Capital, Alexandria Venture Investments and “other key investors.”