The Future of Phenotyping: What’s Next in AI

What’s the next step in the age of AI? (Credit: Pixabay)

Artificial intelligence, increasingly the cornerstone for advancing precision medicine worldwide, is only as good as the data that’s fed to it.

But what’s the next step in the age of AI? The answer may be staring us in the face: it’s your phenotype. We reported previously on a novel approach to phenotyping, facial analysis combined with artificial intelligence, which has been a successful method to aid in identifying difficult-to-diagnose diseases. This cutting-edge approach ultimately shortens the time to diagnosis, gives doctors the ability to spend more time treating patients, and gives drug developers insight into personalized therapeutic approaches.

Facial analysis is a type of next-generation phenotyping (NGP), a technology we recently reviewed in our Clinical Genomics 101, that is revolutionizing the diagnostic process and discovery of new disease. NGP technologies capture, structure, and interpret complex physiological information, and seek to improve our understanding of genetics and its impact on health.

Next-generation sequencing (NGS) is widely known and accepted within the healthcare community, but as NGP continues to come on the scene, it’s increasingly clear that the two go hand-in-hand, as perfect complements, with the only logical progression being that, “next-generation sequencing should lead to next-generation phenotyping,” according to a recent paper published in the Journal of Inherited Metabolic Disease.

Dr. Peter Krawitz, Director of the Institute for Genomic Statistics and Bioinformatics at the University of Bonn in Germany and Chief Data Science Officer at FDNA, is currently leading a team at Bonn to quantify the impact NGP technologies have on NGS data interpretation. The interim results from the Prioritization of Exome Data by Image Analysis (PEDIA) study already show that through combining NGP technology, clinical annotations, and NGS data, the correct diagnosis ranked first 81 percent of the time, compared to the typical 25 percent diagnostic yield of NGS data alone.

“With NGP, we now have the ability to identify the facial phenotype of so many of these traditionally difficult-to-diagnose diseases,” said Ilana Jacqueline, Manager of Patient Advocacy at FDNA. Echoing that sentiment, Dr. Christine Stanley, Head of Clinical Laboratory, US, at leading genomic information company, WuXi NextCODE, went on to say, “The future is not static, it is digital and dynamic. Analyzing the phenotype helps us to define variant prioritization and clinical genetic reporting, ushering us into an era of more personalized genetic testing.”

One of the common issues with phenotyping technologies is a lack of data and understanding of disease phenotypes among ethnic minorities. This gap, while still a reality, is quickly closing through projects that are focusing on global phenomic and genomic data. Recent studies show that “gender and ethnic background do not affect performance” of these NGP technologies.

“There’s been a lot of hype around AI in healthcare,” said Jacqueline, “and it hasn’t always been for the best, but with these new advances we’re starting to see that hype replaced with hope.”