AI May Know How You'll Respond to a Vaccine Before You Get It

Vaccines prevent serious illness for many people, but the immune protection they produce can vary substantially from person to person. New research led by Arizona State University suggests the immune system may show signs of how strongly it will react even before a person is vaccinated.

Researchers at ASU and collaborating institutions examined blood samples from more than 4,000 people, measuring antibodies that recognized 185 antigens, including common viruses and bacteria as well as targets linked to autoimmune diseases. Artificial intelligence was then used to search for patterns in blood samples taken before and after COVID-19 vaccination, uncovering antibody signatures that could help separate people who produced strong vaccine responses from those whose responses were weaker. "What our study found is that certain biomarkers, when analyzed with AI, can predict who is likely to respond well to a vaccine, even before they receive it. This suggests that some people may be more immune-ready than others," said Joshua LaBaer, who led the study and is executive director of the Biodesign Institute at ASU.

The team examined 8,687 samples from 4,089 participants, including healthy volunteers as well as people with conditions or treatments associated with immune suppression, such as HIV, multiple myeloma, solid organ malignancy, autoimmune disease, inflammatory bowel disease and solid organ transplantation. Several immunosuppressed groups were more likely to show reduced responses to COVID-19 vaccination, yet simply placing someone into an immunosuppressed or healthy category did not reliably predict the outcome — some participants with suppressed immune systems still developed strong responses, while about 5% to 6% of healthy participants showed weak vaccine responses.

Certain antibodies already present before vaccination stood out in the analysis. Higher levels of antibodies targeting common microbes, including Staphylococcus aureus, RSV and human respirovirus 3, were associated with stronger responses to COVID-19 vaccines. The researchers call these "sentinel" antibodies, since they may indicate how prepared a person's antibody-producing immune system is to mount a response, even though they are not necessarily acting directly against the vaccine target.

A deep learning model that examined patterns across the entire antibody panel provided more predictive information than a handful of individual biomarkers, the researchers found — suggesting that understanding vaccine readiness may require looking at the immune system as an interconnected whole rather than focusing on one antibody or one disease.