About 10% of people with glaucoma who receive proper treatment still go blind, according to the Glaucoma Research Foundation. Qrativ, a new startup formed by Mayo Clinic and artificial intelligence (AI) technology company nference, wants to give that 10% a chance to keep their vision. It will do it by developing a treatment using a combination of artificial intelligence and pharmaceuticals in their early stages of development.
Rare diseases and diseases with "unmet medical needs," such as those that don't respond to medication, comprise a small but no less deserving segment of the drug population. Of more than 7,000 types of rare diseases, only 5% have an FDA-approved drug treatment, according to EveryLife Foundation for Rare Diseases.
Qrativ founder/CEO Murali Aravamudan said the company hopes to improve this statistic with its AI-derived precision medicine approach.
"A treatment we find may only work for a subset of a cohort," he said. "The objective is to figure out how to maximize the cohort that has an unmet need. If the subset is a reasonable subset, it's still interesting from a commercial standpoint."
The Qrativ Approach
Qrativ formed after Dr. Andrew Badley, Mayo Clinic's director of the Office of Translation, accomplished in two months what his lab had been unable to accomplish in the prior two years by using nference's nferX synthesis platform.
"We had a longstanding data analytics problem we were previously unable to resolve," said Badley, who also serves as Qrativ's cofounder and chief medical officer. "I explained the problem to Murali, and about two months and not much money later, he solved the problem. It was an important discovery in my research. That led Murali and I to discuss how we could synergize our areas on a more global scale."
The result brings together nference's nferX artificial intelligence software platform with Mayo Clinic's physicians, scientists, and clinical data to ultimately conduct new clinical trials on unmet medical needs. AI does the heavy lifting. Humans make sense of the results.
"The AI system doesn't give you one answer. It gives you a list of drug combinations with levels of confidence," Aravamudan explained. "This is where Mayo's clinical expertise comes into play. Andrew's department assembles disease experts to vet the biological and medical rationale. It's the closed-loop of machine telling humans and humans telling back to machine that makes this whole thing happen."
The AI Factor
The machine component, nferX, extracts knowledge from millions of public biomedical documents—abstracts, PubMed papers, clinical trials, FDA records, and SEC filings—using neural networks. From this unstructured text, scientists can ask questions about whether gene A is related to disease B, for example.
"If the relationships are statistically significant, either the association is well-known or the literature doesn’t talk about the association, but the system says the data is associated in a significant way," Aravamudan said. "Those are interesting signals."
To gain confidence about the association, Qrativ researchers overlay structured data sets, such as data culled from Mayo Clinic clinical trials and investigations and other molecular and genomic data sets. "Sometimes we get interesting biological associations that underpin potential new discoveries and science we want to chase," Aravamudan said.
Qrativ works closely with Mayo Clinic to develop and test drugs or drug candidates for the clinic's translation, though it has broader goals, as well. Qrativ partners with pharmaceutical companies, small biotech companies, scientists, and others who have early-stage drugs or drugs in development—Qrativ's market niche—to advance new treatments. " Based on AI, medical knowledge, and the understanding and expertise of the physician-scientists at Mayo, we can determine which disease and which population would be most likely to benefit," Badley said.
If Qrativ decides to advance a product, it assembles what it calls a translation board comprised of researchers, doctors, and other specialists with experience in the field where the drug has promise. The board weighs in on a drug's potential side effects and complications, how to study the drug in clinical trials, and what types of patients to recruit.
Qrativ certainly isn't the first company to use AI for drug research and development. But its focus on rare diseases, combined with its ability to bring in Mayo Clinic expertise, sets it apart. "We're helping people for whom there is no other treatment," Aravamudan said. "That's the basis."