103Despite existing reports on differential DNA methylation in type 2 diabetes (T2D) and obesity, our 104 understanding of the functional relevance of the phenomenon remains limited. Because obesity is the main 105 risk factor for T2D and a driver of methylation from previous study, we aimed to explore the effect of DNA 106 methylation in the early phases of T2D pathology while accounting for body mass index (BMI). We performed 107 a blood-based epigenome-wide association study (EWAS) of fasting glucose and insulin among 4,808 non-108 diabetic European individuals and replicated the findings in an independent sample consisting of 11,750 non-109 diabetic subjects. We integrated blood-based in silico cross-omics databases comprising genomics, 110 epigenomics and transcriptomics collected by BIOS project of the Biobanking and BioMolecular resources 111 Research Infrastructure of the Netherlands (BBMRI-NL), the Meta-Analyses of Glucose and Insulin-related 112 traits Consortium (MAGIC), the DIAbetes Genetics Replication And Meta-analysis (DIAGRAM) consortium, and 113 the tissue-specific Genotype-Tissue Expression (GTEx) project. We identified and replicated nine novel 114 differentially methylated sites in whole blood (P-value < 1.27 × 10 -7 ): sites in LETM1, RBM20, IRS2, MAN2A2 115 genes and 1q25.3 region were associated with fasting insulin; sites in FCRL6, SLAMF1, APOBEC3H genes and 116 15q26.1 region were associated with fasting glucose. The association between SLAMF1, APOBEC3H and 117 15q26.1 methylation sites and glucose emerged only when accounted for BMI. Follow-up in silico cross-omics 118 analyses indicate that the cis-acting meQTLs near SLAMF1 and SLAMF1 expression are involved in glucose level 119 regulation. Moreover, our data suggest that differential methylation in FCRL6 may affect glucose level and the 120 risk of T2D by regulating FCLR6 expression in the liver. In conclusion, the present study provided nine new DNA 121 methylation sites associated with glycemia homeostasis and also provided new insights of glycemia related loci 122 into the genetics, epigenetics and transcriptomics pathways based on the integration of cross-omics data in 123 silico. 124 125 146 Results 147 1. Blood-based epigenome-wide association analysis of glycemic traits 148The discovery phase was based on four European cohorts (Supplementary Table 1). The meta-analysis 149 revealed DNA methylation in 28 unique CpG sites associated with fasting glucose (11 CpG sites, n = 4,808) 150 and/or insulin (20 CpG sites, n = 4,740) at epigenome-wide significance (P-value < 1.27 × 10 -7 ) in either the 151 baseline model without body mass index (BMI) adjustment or in the second model with BMI adjustment. Of 152 these 28 CpG sites, 15 were novel (Table 1) while 13 were identified by earlier EWAS studies of either T2D or 153 related traits, including glucose, insulin, hemoglobin A1c (HbA1c), homeostatic model assessment-insulin