This story has been updated to clarify Coral Genomics' progress on its Series A financing round.
NEW YORK – While personalized treatment approaches continue to amass in oncology, Coral Genomics, a startup founded by scientists who did their graduate studies at the Broad Institute, is betting that opportunities lie in other fields of medicine, too.
The company is generating experimental datasets using samples that capture a large fraction of human genomic diversity. It combines a proprietary sample preparation method for shallow sequencing with functional studies and algorithms to predict response to drug treatment.
"Sample sizes for disease risk and drug response studies are small and often phenotype is quite hard to define," said Atray Dixit, Coral Genomics' CEO and cofounder. "With large-scale, diverse, in vitro models, we can increase the utility of precision medicines for diverse populations as well as start to model drug response at scale."
Coral's approach has started to win it some converts. Last week, the firm partnered with Denmark's Leo Pharma, a company developing products for treating skin conditions, and in March, it partnered with DrugViu, a clinical data company that also focuses on genomic diversity and autoimmune diseases. Coral also received two different Small Business Innovation Research grants from the National Institutes of Health in 2020, totaling approximately $3.6 million.
"The gap they're filling is a really important one," said Jonathan Rosenberg, a gastroenterologist at Illinois Gastroenterology Group, who is working with the company on a clinical study of patient response to treatments for inflammatory bowel disease, or IBD, which includes Crohn's disease and ulcerative colitis. "Right now, we have a wide array of medications to try to target various inflammation pathways to address underlying causes, but we don't have any tools to decide which medication is right for the patients."
"It's not just about picking the right medication — picking the wrong medication has huge downstream effects economically," Rosenberg said.
Coral Genomics started up in a whirlwind over the summer of 2018. While working on his doctorate, which also included work on Perturb-seq, a CRISPR-based phenotype screening method, Dixit met his cofounder, Jacob Borrajo, who is still finishing his doctorate at the Massachusetts Institute of Technology. The pair had waded into the Boston-area startup scene with their idea for shallow sequencing and were able to get a small grant from the MIT Sandbox Innovation Fund to get off the ground. They also applied to the legendary Y Combinator accelerator program, which had launched companies such as Airbnb, Dropbox, and DoorDash. A week before defending his thesis, Dixit was flown out to the Bay Area for an interview and later found out that Coral had been accepted. He started at Y Combinator the same day that he graduated.
The experience helped Coral refine its library prep technology, adopt a clinical focus, and ultimately raise about $2 million in seed funding after its demo day. The company, now based in San Francisco, has a handful of employees, including a CLIA laboratory team and some part-timers, and is looking to add another three or four staffers by the end of the year. Coral CSO Kasey Davis, now also an adjunct professor at Stanford University, joined in 2019 and brings key experience in primary cell assays and recruiting patients, Dixit noted.
According to Dixit, Coral's sequencing sample prep method can reduce sequencing costs by 50 percent. That's potentially huge when clinical studies involve whole-exome and whole-genome sequencing analysis. The method involves sample pooling, but he was cagey about divulging more. "Part of it is trade secret and some of it is covered by a patent that has not yet been issued," he said.
According to the SBIR award, the method is capable of sequencing hundreds of samples in less than a week and, with further development, has the potential to bring the cost of shallow-coverage WES and WGS down to $15 or less per sample. Coral also noted that it has developed a real-time analysis platform, capable of returning variant calls within two hours.
"Shallow sequencing has a lot of utility for essentially replacing microarrays as a way of genotyping," Dixit said, especially as reference panels get more complete and imputation methods get more accurate. However, there is a sweet spot, which depends on the questions one is asking. "You can go surprisingly shallow in coverage, as well as number of cells, to get a lot of information content," Dixit said.
Coral is not looking to commercialize its sample prep method as an independent product but uses it in genotyping samples for its research and clinical studies. The technology has helped in the firm's other SBIR grant, a Phase II award, which funds a study that characterizes response to treatment in IBD as a function of genetic variability in a diversity of populations. The company combines its genotyping data with functional tests of drugs against primary immune cells, with RNA sequencing analysis as an endpoint, often using single-cell methods. Shallow sequencing also has value in RNA-seq: "You get diminishing returns in sequencing lowly expressed genes at high depth," Dixit said.
Coral has accessed samples from some large repositories, including the UK Biobank, BioBank Japan, as well as other data sources. The company is also partnering with DrugViu to help find patients for its IBD study. "They have their own internal efforts to find [IBD] patients who might be interested in participating in research, and they're kind of steering them in our direction," Dixit said. "We're mostly paying them, basically per lead." DrugViu has also helped with interpreting patients' electronic health records and providing data on disease severity and progression after treatment has been administered.
Making a clinical test to predict response to drugs used to treat autoimmune diseases is Coral's main area of focus in the near term. Spending on medications for autoimmune disease is about $54 billion per year, Dixit said, a number that is growing every year.
The company is not alone in its quest, though, even in the IBD space. The Vectra DA test for rheumatoid arthritis, for example, has been available for more than a decade. Myriad Genetics acquired the proteomics-based test in 2014 as part of its $270 million purchase of Crescendo Bioscience but said earlier this week that it was selling the test to Labcorp for $150 million. Also, Scipher Medicine, which recently raised $82 million in Series C financing for rheumatoid arthritis, has already produced a peer-reviewed validation study of its PrismRA RNA-based test to guide therapy decisions. In addition, Scipher has signed a drug development deal with pharma firm Galapagos. Further, PredictImmune, which has developed PredictSure IBD, a 17-gene PCR panel that the firm claims can guide IBD treatment, has signed a multinational commercialization pact with Cambridge Clinical Laboratories.
Coral's initial research cohort for IBD has hundreds of patients, and the firm is testing responsiveness to different drug classes in vitro. "We saw tremendous variations" in response to treatment, Dixit said, highlighting the fact that Coral's studies are reflecting "real-world populations," including an important segment of IBD patients: those who do not respond to any of the available treatments.
A test that could provide information related to the spectrum of treatment options, Rosenberg said, "would not entirely guide my decisions, but would help me have a better-informed discussion with patients up front, setting expectations." For non-responders, the treatment journey can end with surgery to remove the diseased section of bowel, but everything else is thrown up "like darts at a wall in the hope something sticks," he said.
Coral's IBD study will continue until 2022, but the firm hopes to release preliminary results before the end of 2021. In the meantime, Dixit is thinking about putting together the firm's Series A financing round, though he has not officially started yet and did not say how much the company is hoping to raise.