Identification of immune responses predictive of clinical protection against Malaria using novel statistical pipelines

BackgroundNowadays, the chance of discovering the best antibody candidates for explaining naturally acquired protection to malaria and detecting exposure to malaria parasites has notably increased due to publicly available multi-sera data. The analysis of these data is typically divided into a feature selection phase followed by a predictive one where several models are constructed for the outcome of interest. A key question in the analysis is to determine which and how each feature should be included in the predictive stage. ResultsTo answer this question, we developed three approaches for classifying malaria protected and susceptible groups: (i) a basic and simple approach based on selecting antibodies via the nonparametric Mann-Whitney test; (ii) a dichotomization approach where each antibody was selected according to the optimal cut-off via maximization of the {chi}2 statistic for two-way tables; (iii) a hybrid parametric/non-parametric approach that integrates Box-Cox transformation followed by a t-test, together with the use of finite mixture models and the Mann-Whitney test as a last resort. We illustrated the application of these three approaches with published serological data for predicting clinical malaria in 121 Kenyan children. The predictive analysis was based on a Super-Learner where predictions from multiple classifiers were pooled together. Our results led to almost similar areas under the Receiver Operating Characteristic curves of 0.72 (95% CI = [0.61, 0.82]), 0.80 (95% CI = [0.71, 0.90]), 0.79 (95% CI = [0.7, 0.88]) for the simple, dichotomization and hybrid approaches, respectively. ConclusionsThe three feature selection strategies provided a better predictive performance of the outcome when compared to the previous results solely relying on Random Forests alone (AUC=0.68). Given the similar predictive performance, we recommended the three strategies should be used in conjunction in the same data set and selected according to their complexity.