Blending datasets for better research

AI collaboration seeks to improve access to COVID-19 datasets across institutions

February 24, 2021
Mel J. Yeates
Blending datasets for better research

WEST LAFAYETTE, Ind.—During the COVID-19 pandemic, healthcare professionals and researchers have largely been limited to using local and national datasets to study the impact of various factors on the course of the disease.

Now, Purdue University is joining with other organizations for an initiative to accelerate global collaborative research on COVID-19. The project aims to provide access to high-quality, real-time multicenter patient datasets. A sum of $256,000 to develop the Records Evaluation for COVID-19 Emergency Research (RECovER) initiative has been provided by the National Science Foundation.

“Onai is a company with a focus on cryptography, distributed ledger technology, and AI,” says Shriphani Palakodety, software engineer for Onai, a company specializing in blockchain and crypto technology. “With the COVID-19 pandemic, Onai realized that our approaches could help enable more and faster collaborative research efforts, allowing queries to be quickly performed on patient records across institutions in a completely secure manner. We therefore launched RECovER.”

Researchers are testing predictions of artificial intelligence drug discovery platforms from the lab of Prof. Gaurav Chopra, an assistant professor of analytical and physical chemistry in Purdue’s College of Science, on patient datasets across a network of health care institutions. The team is using the technology to find trends and data connections to help understand and treat COVID-19, with an emphasis on the impact existing medications have on the virus.

“The Chopra laboratory has been developing machine learning and modeling methods for drug repurposing and drug design. We [previously] published a paper in Drug Discovery Today on drug repurposing for pandemics,” notes Chopra. This collaboration makes it possible to do decentralized electronic health record dataset searches from several medical institutions.”

“We are happy to collaborate with several academic researchers, including Prof. Gaurav Chopra, who is a world leader in drug repurposing. There were five organizations involved to start, and we are welcoming participation from additional medical centers to join this effort,” Palakodety tells DDN.

Other institutions involved in the initiative include SUNY Buffalo, the Cincinnati Children’s Hospital Medical Center, and the University of Cincinnati College of Medicine.

“Onai is pleased to enable rapid secure research across COVID-19 medical records,” added Galana Gebisa, a manager at Onai. “Applying Onai technology, data remains local and private within each participating institution. By querying across institutions, researchers can access a larger, more statistically significant dataset, to learn useful information earlier than they otherwise would.”

“Traditionally, researchers have been confined to using local and/or manually assembled datasets to study the impact of comorbidities, pre-existing medication use, demographics, and various interventions on disease course. With RECovER, we can examine all these things in a more scalable and secure manner,” Palakodety explains. “This includes retrospectively examining drug predictions on patient data.”

According to Chopra, “Our approach will allow healthcare professionals or researchers to identify patterns in patient responses to drugs, select or rank the predictions from our platform for drug repurposing, and evaluate their responses over time. Getting patterns from such responses in different diseases and correlating to the proteome scale interactions models for specific response, such as a context-dependent immune response, will allow for added confidence in addition to the known mechanism of the drug candidates at hand. We believe such an approach will help with COVID-19 and other potential pandemics.”

“In short, we want to accelerate cross-institutional research, without any sensitive information being revealed from each institution.The more institutions that join, the larger the set of COVID-19 data, and the more that researchers can learn in a statistically rigorous way,” Palakodety reports. “We think this application of novel computational technology to medical research opens many doors, even looking beyond the pandemic. Compared to the status quo, it allows the community to improve research and privacy at the same time.”

“Our long-term goal is to not only repurpose drugs, but also design new ones in the context of their proteomic interaction environments in a faster, cheaper, safer, and better manner,” commented Ram Samudrala, a senior co-author of the Drug Discovery Today paper.

“The marriage of AI with data security is a no-brainer for solving hard healthcare problems and our project attempts to take this first step. This will streamline efforts for data security and rapid response in the event of current or future pandemic,” Chopra concludes.

Subscribe to Newsletter
Subscribe to our eNewsletters

Stay connected with all the latest DDN news!

Latest Issue  

• Volume 17 • Issue 5 • May 2021

May 2021

May 2021 issue