NEW YORK — Artificial intelligence, or AI, and metagenomics may have a large impact on the future of personalized medicine and precision diagnostics in particular, according to Matthew Might, PhD, professor of medicine and director of the Hugh Kaul Precision Medicine Institute at the University of Alabama at Birmingham.
“Personalized medicine will increasingly exploit AI to deliver truly tailored care to individual patients,” Might told Infectious Diseases in Children. “I hope clinicians take away that metagenomics can be used creatively to help identify pathogens and combined with advances in AI.”
Might, whose professional background is primarily in computer science and cyber security and includes work with the Obama administration’s Precision Medicine Initiative, entered the precision medicine field when his son Bertrand was diagnosed with the first ever case of NGLY1 deficiency — a rare disorder that causes developmental delay, seizures and the inability to produce tears.
"Precision medicine is really ultimately about optimizing health with data," Might said during a presentation at the Infectious Diseases in Children New York Symposium.
During the presentation, Might discussed how the clinical accessibility of the human genome sparked the advancement of AI as a tool for personalized medicine.
“We're trying to find how to modulate parameters, like treatment and lifestyle, to find the optimal health for that particular patient,” Might said. “That's really the goal — to go all the way from the data into the treatment.”
Precision medicine consists of two necessary halves — precision diagnostics to identify the molecular cause of the disease, and precision therapeutics, according to Might. Metagenomics is useful for revealing the cause of an infection, whereas AI is the “big hammer” of precision medicine that can be a useful tool for diagnoses, he said.
After Bertrand’s diagnosis, Might began seeking out potential treatments using various AI tools, despite being told there were no current treatments.
“You can always do science, when there's nothing else you can do,” Might said.
Inferring Bertrand’s inability to develop tears was a consequence of a missing metabolite, Might conducted a natural history study with the NIH to further explore NGLY1, eventually leading to a University of Utah grant to explore therapeutics for potential drug targets. Using planaria — a flatworm — Might discovered Prevacid (lansoprazole, GlaxoSmithKline) was an effective ENGase inhibitor, which aided in Bertrand’s treatment.
Might’s developments eventually led to his role as director of the Hugh Kaul Precision Medicine Institute, where he works to aid children in positions similar to his son’s.
"The goal was really to scale up the kinds of things we'd done for Bertrand for as many patients as possible," Might said. “Really the key here is drug repurposing — finding existing drugs that in some way target a novel disorder."
Exploring precision medicine also led Might to develop mediKanren, an AI diagnostic tool that combines “old-school logic” with “high-speed reasoning” to achieve “superhuman-like deduction with clinical insight,” he said.
"Precision ID is possible,” Might said. “We demonstrated it can be done — there's actually a lot more I think that we can do here — and I really do think that a fusion of AI and metagenomics holds a lot of potential going forward." – by Eamon Dreisbach
Might M. Personalized medicine: the future is now. Presented at: Infectious Diseases in Children Symposium. Nov. 23-24; New York.
Disclosure: Might reports no relevant financial disclosures.