Feature selection and causal analysis for microbiome studies in the presence of confounding using standardization.