AI Identifies COVID-19 Molecular Mimics Linked to Autoimmune Disease Risk

COVID-19 infection has been linked to a heightened risk of autoimmune disorders, including rheumatoid arthritis and type 1 diabetes. Yet, the exact mechanisms by which the virus disrupts the immune system remain unclear—posing a challenge for developing therapies aimed at preventing post-COVID autoimmunity.

One leading hypothesis involves viral “molecular mimics”—proteins from the virus that resemble the body’s own proteins. These mimics may trigger an immune response against the virus but unintentionally cause the immune system to target healthy tissues as well.

Thanks to recent advancements in data analysis and machine learning, scientists have now identified a set of SARS-CoV-2-derived molecular mimics that may play a role in initiating autoimmune responses. The findings were published in the journal ImmunoInformatics.

To begin, researchers looked for viral components that closely resemble human proteins commonly associated with autoimmune diseases. The theory: these viral lookalikes might mislead the immune system into attacking the body’s own cells.

They then used machine learning models to narrow down the list to viral elements most likely to bind with human antibodies—those with the highest potential to trigger an autoimmune reaction.

Among the components flagged, some have been linked to autoimmune conditions like type 1 diabetes and multiple sclerosis. Interestingly, several of the human proteins identified as likely targets appear only in people with certain genetic backgrounds. This suggests that individuals who produce these specific proteins may be at greater risk of developing COVID-related autoimmunity.

“It’s exciting that, in collaboration with our clinical colleagues, we can now use AI and machine learning to tackle medical issues made worse by the COVID pandemic,” said Julio Facelli, PhD, Distinguished Professor and senior author of the study at the University of Utah Health’s Department of Biomedical Informatics. “We hope these findings will pave the way toward better understanding, treatment, and prevention of these debilitating conditions.”