Arizona State University researchers, examining more than 8,600 blood samples from over 4,000 people, have identified antibody signatures already present before vaccination that flag “immune readiness” — including some healthy people who turn out to be weak responders, and some immunosuppressed patients who respond surprisingly well.
Why some people mount a robust immune response to a vaccine while others respond only weakly has remained a persistent and clinically significant puzzle in immunology, particularly for the growing population of patients on immunosuppressive treatment or living with conditions that compromise immune function. A study led by Joshua LaBaer’s team at Arizona State University’s Biodesign Institute, first reported by ASU News on 20 August 2026 and covered widely since, including by ScienceDaily, GEN and Mirage News through 21–22 August, offers a new and potentially clinically translatable approach: using artificial intelligence to read pre-vaccination antibody patterns as a predictor of how strongly a person’s immune system will respond once vaccinated.
What the researchers measured
The team examined 8,687 blood samples from 4,089 participants — a mix of healthy volunteers and people with conditions or treatments associated with immune suppression, including HIV, multiple myeloma, solid organ malignancy, autoimmune disease, inflammatory bowel disease and solid organ transplantation. For each sample, researchers measured antibody responses against 185 different antigens: targets drawn from SARS-CoV-2, a range of other common viruses and bacteria, and markers associated with autoimmune conditions. Rather than the conventional approach of measuring vaccine response only after the shot is given, the team asked whether the antibody “fingerprint” already present in a person’s blood — before any vaccination — could itself signal how strongly that person’s immune system was primed to respond.
Using machine learning to analyse these broad antibody patterns, the researchers identified what they term “sentinel” antibodies: pre-existing antibody signatures that correlated with subsequent COVID-19 vaccine response strength. Notably, the pattern was not simply a proxy for overall health status — some healthy volunteers with no diagnosed immune-compromising condition nonetheless showed signs of blunted vaccine response, while some patients with immunosuppressive diagnoses mounted unexpectedly strong responses, suggesting the antibody fingerprint captures something more specific than a general health or disease-status signal.
Why this approach is methodologically distinct
Existing strategies for predicting vaccine response have generally relied on genetic testing or on measuring gene-expression signatures in blood immediately around the time of vaccination — approaches that are informative but not always straightforward to deploy in routine clinical settings. The ASU-led approach instead analyses antibody patterns already present in blood drawn before vaccination, a measurement type that is comparatively easier to standardise and translate into clinical practice, since it does not require genetic sequencing infrastructure or precisely timed peri-vaccination sampling. The researchers describe their method as one of the first to use this kind of broad, pre-vaccine antibody fingerprint as a measure of what they call “immune readiness.”
Why it matters
If validated in further, larger and more diverse cohorts, this kind of predictive tool could allow clinicians to identify, before vaccination, which patients — particularly among the substantial and growing population on immunosuppressive therapy for cancer, autoimmune disease or organ transplantation — are likely to mount an inadequate response, and to act accordingly: through additional vaccine doses, alternative vaccine formulations, or closer post-vaccination monitoring. This has direct relevance beyond the COVID-19 context in which the current study was conducted, given the broader and growing use of immunosuppressive biologics across oncology, rheumatology and transplant medicine worldwide, including in India, where a substantial and expanding patient population relies on such therapies. The caveats are those typical of a single-cohort predictive-biomarker study: the antibody-response analysis was validated against COVID-19 vaccination specifically, and it remains to be established how well the same “sentinel” antibody patterns generalise to other vaccine types, other populations, and prospective (rather than the retrospective correlational) clinical use. As with any AI-derived biomarker signature, independent replication in separate cohorts — ideally including populations outside the study’s original demographic base — will be an essential next step before this becomes a tool clinicians can act on.
-Dr Srinayana Kavuri
Key facts
– Study: 8,687 blood samples from 4,089 participants (healthy volunteers plus people with HIV, multiple myeloma, solid organ malignancy, autoimmune disease, IBD, or solid organ transplantation)
– Measured antibody responses to 185 antigens (SARS-CoV-2, other common pathogens, autoimmune-disease markers) using AI/machine learning analysis
– Identified pre-vaccination “sentinel” antibody signatures correlating with subsequent vaccine response strength
– Some healthy volunteers showed blunted responses; some immunosuppressed patients responded strongly — the signature is not simply a health-status proxy
– Led by Joshua LaBaer, Arizona State University Biodesign Institute; first reported 20 August 2026
