MACHINE Learning will help scientists understand how COVID-19 vaccines elicit different immune responses in people infected with HIV, providing new insights that can support more personalized vaccination strategies.
Vaccines work by training the immune system to recognize and fight infections. However, immune responses vary widely from person to person, especially in people with underlying conditions such as HIV that affect immune function. Understanding these differences is important for vaccines to provide strong and durable protection to vulnerable groups.
In the new study, researchers used machine learning models to analyze complex immunity data from older adults who had received up to five doses of the SARS-CoV-2 vaccine. Participants included HIV-infected individuals on antiretroviral therapy and HIV-negative individuals of similar age.
Machine learning identifies unique immune characteristics
The research team used a random forest machine learning approach to examine multiple immune markers, including cytokines, antibodies, and indicators of cellular immune activation. This analysis revealed distinct immune signatures that differentiate vaccinated HIV-infected and HIV-negative participants.
One key finding was that the combination of cytokines and saliva-based antibodies was particularly effective at differentiating immune responses in HIV-infected individuals. These signals likely reflect differences in T cell activation and mucosal immunity, two important components of the body’s defense against viral infections.
Interestingly, traditional blood antibody measurements alone were not very informative in identifying these differences. This suggests that saliva-based immune markers may provide further insight into vaccine responses in immunocompromised populations.
Evidence of immune recovery in some patients
Machine learning analysis also revealed promising results. Some HIV-infected participants showed a pattern of immune response after vaccination that was very similar to that of HIV-negative participants. Researchers interpret this as evidence that effective antiretroviral therapy may restore aspects of immune function sufficiently to cause a near-normal vaccine response in some people.
Visualization techniques, including dimensionality reduction mapping, further highlighted clusters of immune features associated with different patterns of humoral and cell-mediated immunity.
Synthetic data and the future of precision vaccination
To expand the potential applications of this study, the researchers generated privacy-preserving “virtual patients” that reproduced the patterns seen in the original immune data. Machine learning models trained on these synthetic datasets can accurately classify immune responses in real patients, suggesting a promising approach for research while protecting sensitive health information.
Overall, this discovery shows how machine learning can help decipher complex vaccine immunology. By identifying key immune characteristics that shape vaccine responses, this approach could support more targeted monitoring and precision vaccination strategies for people living with HIV and other immunologically vulnerable populations.
reference
Korosek CS et al. Long-term immune profile modeling reveals distinct immunogenicity signatures after five COVID-19 vaccinations in HIV-infected individuals. pattern. 2026; DOI 10.1016/j.patter.2025.101474.
Featured image: Yura from Adobe Stock
