This story is part of an ongoing series on A.I. in healthcare.
Artificial intelligence systems have become a trendy, but useful means of extracting meaning out of the troves of scientific research and clinical data that institutions produce every year. Startups from San Francisco to Boston have launched with the goal of helping pharmaceutical companies decide what drugs to develop.
Now, the nonprofit research group Mayo Clinic is entering the fray by launching a startup with Cambridge, MA-based Nference, an artificial intelligence software maker, that will aim to better match drugs with diseases they might be able to treat. The new company, called Qrativ, announced an $8.3 million Series A round of funding Friday from Matrix Capital Management, Matrix Partners, and Rochester, MN-based Mayo Clinic Ventures.
 |
Want more AI, machine learning, and drug discovery content? Network and engage with industry leaders online this June. Learn more. |
Murali Aravamudan, a longtime technology executive who sold voice-based video search company Veveo to search company Rovi in 2014, is the CEO of Nference and will also be the CEO of Qrativ. Qrativ will use the artificial intelligence software that Nference currently offers to pharmaceutical companies to help them analyze data and clinical research studies in their drug development work. Nference’s operations will continue independently, Aravamudan says.
Mayo is providing Qrativ with the troves of its proprietary data from the various clinical trials and investigations it has performed over the years, for analysis by Qrativ’s artificial intelligence software. The company’s software will also analyze publicly available scientific research, and compare it to the Mayo Clinic data, says Andrew Badley, co-founder and chief medical officer of Qrativ.
The goal will be to use the artificial intelligence tools to examine drugs that are already under consideration as treatments for particular medical conditions—for example, a kinase inhibitor that a researcher is studying as a potential way to affect a specific biological pathway—to see if there might be other pathways where it would be more effective. This might maximize the use of that drug, Badley says.
Mayo Clinic becomes most involved after the A.I. system identifies a group of potential uses for a drug. Badley, who is also the director of an office at Mayo Clinic that says it facilitates the translation of drugs into clinical practice, would help assemble a team of researchers with expertise in the medical field where the drug has promise, according to the company. Those “translation boards” of clinical experts, as Qrativ is calling them, will then help determine how the drug should be studied in clinical trials, and what kind of patients they might recruit.
“The fact that there is clinical data available and that it goes right into the system, that is used to inform the selection process,” Aravamudan says. “We close the feedback loop with the information from the experts. That feedback is then learned by the A.I. system.”
Qrativ believes that process will help drugs advance to Phase 1 or Phase 2 clinical trials. The company could then sell or license its discoveries to pharmaceutical companies. Qrativ won’t commercialize any drugs itself. In terms of medical conditions Qrativ might pursue, it is casting a wide net. There are already a few promising candidates, Aravamudan says, but he declined to name them.
Qrativ will also make money by doing the same work for pharmaceutical companies that want to figure out if their existing compounds could be better used in another way, Aravamudan says.
The list of companies using some sort of A.I. for drug development, let alone healthcare as a whole, is extensive. (Xconomy has been following the development in the sector with a series on A.I. in healthcare.)
In 2015, Palo Alto, CA-based TwoXAR raised a $3.4 million seed round from Andreessen Horowitz for its technology that scours public and proprietary datasets, looking for ways to match drugs with diseases. Another Bay Area company, Atomwise, is aiming to help researchers on some of the earlier stages of the drug development process. Meanwhile, larger companies such as IBM and GE are racing to become leaders in A.I. and healthcare with efforts such as Watson Health and partnerships with research institutions.
Those are only a few examples from a list growing so rapidly that it might be necessary to develop an algorithm to keep track. Qrativ is unperturbed by the competition. The company believes the addition of the human, clinical expertise to its A.I. data analysis helps to differentiate it from the pack.
“We feel we have what it takes given the correct combination of A.I. with the actual clinical translation support,” Badley says. “We maximize it with a human approach.”
For a nonprofit like Mayo, the investment in a business comes with specific restrictions on what it can do with any profits that may result. Any money earned from the deal, which was negotiated by Mayo Clinic Ventures, will go into Mayo’s “nonprofit mission of practice, education, and research,” Badley says.
The deal is, for now, limited to using clinical data from Mayo Clinic, Aravamudan says. Even though he admits that having more data is advantageous, Aravamudan says he doesn’t expect to add data from any other research institutions, other than the publicly available scientific research. Badley says he believes those two components will be enough to inform the A.I. system.
“The goal of this particular relationship is to really help Mayo clinic and Nference,” Aravamudan says. “Ultimately everything is a business, and we will do what makes sense. But the objective is to prove that we can be very effective in getting into the clinic with Mayo as the single strategic partner.”
