Improved inference for vaccine-induced immune responses via shape-constrained methods

We study the performance of shape-constrained methods for evaluating immune response profiles from early-phase vaccine trials. The motivating problem this work involves quantifying and comparing IgG binding responses to first second variable loops (V1V2 region) arising in HVTN 097 100 HIV consider unimodal log-concave compare two vaccines, which is reasonable because data support that underlying densities could have these shapes. To end, we develop novel tests stochastic dominance plug-in estimators squared Hellinger distance between densities. Our techniques are either tuning parameter free, or rely on only one parameter, but their better (the dominance) comparable with nonparametric distance). minimal dependence parameters especially desirable clinical contexts where analyses must be prespecified reproducible. supported by theoretical results simulation studies.