Fig. 6.

Comparison of prediction performance with and without using covariates. The performance of two classifiers: LR and LRCV were compared in predicting the asthma outcome of children at year 3 before and after using covariates together with using the miRNA and mRNA (a) and miRNA and microbiome (b), respectively. The missing values of miRNA, mRNA data were imputed using the median values. w/covariates: including covariates in prediction; w/o covariates: without using covariates in prediction. (c): Correlation between the coefficient of each covariate together with two omics combinations: (1) miRNA and mRNA and (2) miRNA and microbiome in the prediction of asthma status using Logistic Regression. 1: History of asthma in mother; 2: History of asthma in father. 3: History of hay fever or allergic rhinitis in mother; 4: History of eczema or atopic dermatitis in mother; 5: History of eczema or atopic dermatitis in father; 6: History of hay fever or allergic rhinitis in mother; 7: Marital status of mother; 8: Highest level of school mother has completed; 9: Ethnicity of mother; 10: Race of mother; 11: Enrollment site; 12: assignment; 13: Treatment Arm; 14: Vitamin D blood value (ng/ml) at Enrollment visit; 15: Mother's age at enrollment; 16: Mother's gestation age in days, at enrollment; 17: Mother's Education re-worked; 18: Mother's Marital status; 19: Race/Ethnicity