Abstract
MOTIVATION: Eosinophils are phagocytic white blood cells with a variety of roles in the immune system. In situations where actual counts are not available, high quality approximations of their cell proportions using indirect markers are critical. RESULTS: We develop a Bayesian measurement error model to estimate proportions of eosinophils in cord blood, using the cord blood DNA methylation profiles, based on markers of eosinophil cell heterogeneity in blood of adults. The proposed method can be directly extended to other cells across different reference panels. We demonstrate the method's estimation accuracy using B cells and show that the findings support the proposed approach. The method has been incorporated into the estimateCellCounts function in the minfi package to estimate eosinophil cells proportions in cord blood. AVAILABILITY: estimateCellCounts function is implemented and available in Bioconductor package minfi. SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.