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

Abstract

AbstractIdentifying antibody responses associated with natural immunity to malaria is key to advancing antimalarial vaccine development. With the advent of high-throughput serological assay robust pipelines that produce solid and reproducible results are crucial for the identification of the antibody responses that lead to malaria protection.Here we have developed two pipelines and assessed their predictive performance on published data from IgG antibody responses against 36 antigens derived from Plasmodium falciparum in 121 Kenyan children with ages below 10 years old. The first pipeline relied on parametric approaches while the second represented a more pragmatic approach to data analysis.The proposed pipelines enabled us to construct classifiers based on few antibodies, whose performances outperformed previous findings based on Random Forest. The best classifier overall was based on antibodies against the msp2, msp4, msp7, msp10, pf11_0373 and pf113 antigens and reached a predictive performance of 86% (AUC = 0.86; 95% CI = (0.79-0.93)) using the pragmatic approach. Concerning the parametric approach, our best achievement was a classifier against the h103, msp2, msp4, msp7 and msrp3 antigens with a predictive performance of 82% (AUC = 0.82; 95% CI = (0.75-0.90)). The good performance of our pipelines suggests their applicability in antibody data analysis intending to identify antimalarial vaccine candidates.In summary, we were able to identify several antibody responses with high predictive ability against clinical malaria. The proposed pipelines also showed promise to improve the statistical analysis of antibody data aiming to identify antimalarial vaccine candidates.