NIAID CIO: AI Expands Multidimensional Biomedical Data Analysis

While artificial intelligence is often touted as a society-changing capability, it is one of several tools working to advance biomedical research at the National Institute of Allergy and Infectious Diseases.

“The major lure of AI for me is the multidimensional analysis of the multidimensional data,” said Mike Tartakovsky, CIO and director of the Office of Cyber Infrastructure and Computational Biology at NIAID, during the Health IT Summit Wednesday.

Tartakovsky said NIAID launched a program 11 years ago to collect biomedical data across areas including genomics, clinical research and the microbiome. AI tools are now expanding the agency’s ability to analyze those datasets across multiple domains.

“Analysis of this in the past was done primarily [as], ‘Hey, if I am a clinician, I’ll be more interested in the clinical data, if I’m a radiologist, I’ll be more interested in the imaging data,’” he said. “AI allows you to look in conjunction across all of these domains and finding the interesting points that were previously not looked at or even not considered.”

Tartakovsky said integrating analysis across those datasets could generate new hypotheses to guide future research. The agency has also developed tools that can convert unstructured data into structured formats, further strengthening its research and analysis capabilities.

While Tartakovsky said the number of commercial tools that address NIAID’s specific research needs is currently limited, he expects more AI-enabled capabilities to become available in the future.

Ultimately, Tartakovsky said AI will work alongside other technologies to augment the agency’s research capabilities. He pointed to AI-assisted modeling and virtual reality as examples of technologies NIAID is using to support its investments in advanced microscopy research.

“When we are talking about scientific data, health care data, this is a multidimensional topic. AI is very helpful in analysis, but I don’t want to forget about the other tools that are now available to us,” he said. “AI is applicable to virtually everything that we are doing today, but it’s not the only main driver of the scientific research today.”