University of Pennsylvania researchers called on computational systems biology expertise at Oak Ridge National Laboratory to analyze large datasets of single-cell RNA sequencing from skin samples afflicted with atopic dermatitis. The team, led by UPenn’s Dana Graves and John Seykora and ORNL’s Daniel Jacobson, revealed new insights into the role certain genes play in skin diseases such as eczema.

Jacobson said the study presented an opportunity to use a new explainable-AI based method on real-world health-related datasets on ORNL’s Summit supercomputer. They analyzed UPenn’s skin cell data from mice, plus publicly available human tissue data, with a machine learning technique to compare gene expression through hundreds of millions of possibilities and identified gene expression patterns seemingly responsible for inflammation early in the disease’s progression.

“Our goal is to further refine these studies and extend them,” said Seykora. “By teaming with ORNL we found new genes and cells that may be therapeutic targets for treatment,” said Graves.