Using statistical network methods to elucidate the multi-omic modulators of the effect of maternal vitamin D levels on childhood asthma

Asthma is the most common chronic disease among children globally. Disruptions of fetal development processes are hypothesized to give rise to asthma, therefore, prenatal interventions may help reduce its short- and long-term burden. Randomized clinical trials, including our Vitamin D Antenatal Asthma Reduction Trial (VDAART), have suggested that maternal vitamin D supplementation can help prevent asthma in the offspring. However, in these trials, not all children benefited from maternal vitamin D supplementation, and not all treated mothers had sufficient vitamin D levels. This variability in vitamin D response suggests complex mechanisms underlying the potential protective effect of maternal vitamin D on childhood asthma that are currently not well-understood. Our long-term goal is to uncover the molecular mechanisms that can enable the precision prevention of childhood asthma. The objective of this application is to identify the multi-omic determinants that modulate the influence of maternal vitamin D supplementation on childhood asthma outcomes. Our central hypothesis is that the effect of maternal vitamin D intake on offspring asthma status is modified by the genotypes of and the epigenetic modifications on the major regulators of vitamin D metabolism and signaling. In Aim 1, we will determine the joint role of vitamin D binding protein (DBP) levels and vitamin D metabolism and signaling genotypes in modulating the effect of maternal vitamin D supplementation in childhood asthma outcomes. We will measure DBP levels in the VDAART cohort and use the existing VDAART genotyping data. We will use Mixed Graphical Models and Conditional Gaussian Bayesian Networks to model the contribution of DBP levels, 25-hydroxyvitamin D (25-OHD) levels and vitamin D genotypes to the effect of maternal vitamin D intake on childhood asthma, and create polygenic and polyexposure scores for individual risk prediction. In Aim 2, we will identify the longitudinal epigenetic markers that modulate the effect of maternal vitamin D on childhood asthma. We will measure DNA methylation on VDAART mothers pre- and post-vitamin D treatment and use existing cord blood methylation data. We will perform epigenome-wide association studies, integrate their results with ChIP-seq and ATAC-seq data, and perform mediation analyses to identify the longitudinal maternal epigenetic marks that potentially mediate the effect of maternal vitamin D intake in childhood asthma. In Aim 3, we will determine the pharmacogenomic drivers of individual differences in maternal vitamin D response and its effect on childhood asthma. We will integrate genomics, transcriptomics and epigenomics from VDAART to identify the vitamin D-responsive pharmacogenomic expression and methylation quantitative trait loci (PGx-eQTLs and PGx-mQTLs). We will develop a network-based statistical method to identify the colocalization of these PGx-QTLs, perform colocalization using orthogonal approaches, and identify endotypes through which maternal vitamin D targets childhood asthma. Within each of the above aims, we will leverage the strong non-white representation among VDAART participants to address health disparities in vitamin D response and asthma outcomes across the population. Our research plan is innovative in its use of highly granular multi-omic data and cutting-edge integrative methods, and the outcomes of each aim are of high translational potential.