Responses associated with vaccine recipients

Comparing placebo and vaccine recipients

Comparisons performed using the Wilcoxon–Mann–Whitney test followed by multiple test adjustment using false discovery rate (FDR).

Antibody binding responses

Fc effector functions

Comprehensive vaccine associated profile

Integrating antibody binding and functional responses at peak immunity to define a vaccine profile. First, Random Forest (RF) was used to downselect binding response. The RF-selected features were integrsted with Fc effector functions to define the vaccine profile using Partial Least Squares Discriminant Analysis (PLS-DA) analysis.

Model evaluation was performed for both RF analysis using cross-validation and PLSDA by comparing true model to group permutated model ran 200 random times.

Correlation network between binding and functional responses was created with informative features selected by RF together with significantly correlated features (cor > abs(0.9))

Vaccine induced response durability

Comparisons performed using the Wilcoxon signed-rank test followed by multiple test adjustment using false discovery rate (FDR).

Pairwise correlations among antibody binding responses

Spearman correlation used to evaluate pairwise correlations.

Responses associated with delayed viral rebound

Comprehensive vaccine associated profile

Integrating antibody binding and functional responses at peak immunity to define a vaccine profile. First, Random Forest (RF) was used to downselect binding response. The RF-selected features were integrsted with Fc effector functions to define the vaccine profile using Partial Least Squares Discriminant Analysis (PLS-DA) analysis.

Model evaluation was performed for both RF analysis using cross-validation and PLSDA by comparing true model to group permutated model ran 200 random times.

Example shown for responses at the peak vaccine immunity time point

Comparing ADCP responses by viral rebound group

Comparisons performed using the Wilcoxon–Mann–Whitney test.