The investigators followed a cohort of 381 Large White pigs to test whether measurements at weaning predict later vaccine antibody outcomes. Animals were allocated to two parallel vaccination trials: one targeting Mycoplasma hyopneumoniae (MH; n=189) and the other targeting Influenza A virus (IAV; n=192). Vaccination occurred at 28 days of age with a booster given 21 days later. Long-term antibody (Ab) status was assessed at 118 days post-vaccination (dpv) using assay-specific thresholds defined by the study team. Ab persistence was calculated as the decline from the individual peak response—measured at either 28 or 35 dpv—down to the level observed at 118 dpv. The authors reported that the source data used a wide panel of baseline immune and physiological variables collected at weaning; specific assays and exact panels were described in the original manuscript but are not restated here in full by the source.
Data-driven selection was performed with sparse Partial Least Squares Discriminant Analysis (sPLS-DA) to identify candidate predictors, followed by PLS-DA classification models built on those variables. The study team reported cross-validated classification accuracies between 64.7% and 67.7%, reflecting moderate predictive performance rather than highly discriminative models. The authors presented analyses separately for long-term Ab status and for Ab persistence, and they compared predictor sets across the two different vaccines to determine shared versus vaccine-specific signals. The source did not provide additional methodological details in this summary about parameter tuning, class balance handling, or exact cross-validation scheme; those details are available in the full preprint.
Across both vaccine trials, the authors identified eight predictors common to positive long-term Ab status. Four blood-derived traits were repeatedly associated with a positive outcome: hematocrit, IgG concentration, platelet count, and plateletcrit (PCT). These traits correlated with sustained serological positivity at 118 dpv for both MH and IAV in the reported analyses. Conversely, two hematology measures were associated with negative long-term status across vaccines: mean corpuscular hemoglobin concentration (MCHC) and neutrophil percentage. The study team interpreted these associations as evidence that broad baseline features of blood composition and humoral baseline (IgG) measured at weaning relate to later serological status following vaccination. The source emphasizes that predictors were shared but does not claim that any single marker alone offers definitive prognostic power given the model accuracies.
When the outcome was Ab persistence—the extent to which peak responses were maintained to 118 dpv—twelve predictors were shared across vaccines. Higher vitamin D and α-1 globulin concentration at weaning were associated with more persistent responses in both MH and IAV studies. Platelet count and PCT again appeared among the consistent positive markers for persistence. In contrast, MCHC and several spectral absorbance traits measured at weaning were associated with non-persistent trajectories. The authors reported these convergent predictors across two antigenic contexts, suggesting a mix of metabolic, protein, and hematologic baseline features that track with how antibody levels decline after their peak. The source text did not enumerate every spectral trait or provide mechanistic testing of these associations within the summary.
The study team concluded that immunological and physiological traits measured at weaning are associated with later variability in vaccine-induced antibody outcomes. The findings indicate both shared markers—those that reflect general baseline immunocompetence—and vaccine- or phenotype-specific predictors that differ by antigen and by whether the outcome is binary long-term status or decline-based persistence. Reported cross-validated accuracies (64.7%–67.7%) imply moderate classification ability; the authors therefore framed results as associative and predictive at a population level rather than individually decisive.
Limitations reported or implied in the source include that this work is a preprint and has not undergone peer review. The summary did not provide comprehensive methodological parameters in this abstract-level text—details such as full variable lists, assay thresholds, exact cross-validation folds, and potential confounder adjustment are described in the full manuscript but were not reiterated in the source abstract. The summary also does not claim causality, and it does not report external validation in independent herds or different breeds. Key unanswered questions that remain and that the authors themselves leave open include whether these markers generalize to other vaccines, whether interventions at weaning could modify long-term response, and how much of the observed signal is genetic, environmental, or management related. The preprint invites further validation and peer review to establish reproducibility and potential application in vaccination strategies or selective management practices.