Background: Major depressive disorder (MDD) is primarily treated with antidepressants, yet many patients fail to respond to adequate trials. Understanding who is likely to respond to antidepressant treatment and/or what mediates this response is of considerable clinical importance. As part of the Canadian Biomarker Integration Network in Depression (CAN-BIND-1) initiative, we aimed to identify differential DNA methylation marks as epigenetic predictors of antidepressant response (ADR) in MDD patients. Methods: Healthy participants (n=112) and depressed participants (n=211) between 18-60 years of age were recruited across six Canadian clinical centers. Eligible depressed patients with MDD by DSM-IV-TR criteria and a Montgomery-Åsberg Depression Rating Scale (MADRS) score of ≥24 were enrolled. Genome-wide DNA methylation analysis was conducted using the Infinium MethylationEPIC Beadchipb with DNA extracted from baseline peripheral blood samples prior to beginning an eight-week trial of escitalopram. Genome-wide mRNA expression analysis was conducted on the HumanHT-12 v4 Expression Beadchip in RNA extracted from leukocytes at baseline. Depressed patients were classified as non-responders (NRES) and responders (RES) according to changes in MADRS scores following eight weeks of treatment. Differentially methylated positions (DMPs) were identified in regions of differentially expressed genes and validated using a targeted sequencing approach. Replication was conducted with patients participating in a similar trial, the Douglas Biomarker Study. CAN-BIND-1 clinical trial was registered with the ClinicalTrials.gov identification #: NCT101655706. Findings: After depressed participants concluded the 8-week trial, 82 RES and 95 NRES were included in this study. Genome-wide differential DNA methylation revealed 2,572 DMPs (p<0.05, with FDR = 0.1). 303 DMPs were located within 271 genomic regions after applying a cut-off of two percent absolute change in β values. Differential expression of these genomic regions was assessed (p<0.05, FDR=0.1, logFC ≥0.1). Three DMPs in CHN2 (cg23687322, p = 0.00043 and cg06926818, p= 0.0014) and JAK2 (cg08339825, p=0.00021) gene regions were the most significantly associated with mRNA expression changes and validated with targeted sequencing. One CHN2 site (cg06926818) was successfully replicated in the Douglas Biomarker Study Cohort. Interpretation: DMPs found within CHN2 and JAK2 gene regions may act as predictors of ADR. Interestingly, both genes have some relevance to current hypotheses of MDD etiology and ADR. JAK2 encodes for a tyrosine kinase involved in specific cytokine signaling that mediates peripheral inflammation and CHN2 codes for a GTP-ase activating protein involved in controlling axon pruning processes during neurodevelopment. Although our findings are promising, further studies are required to add clinical validity to our results. Trial Registration Number: ClinicalTrials.gov identification #: NCT101655706. Funding: Canadian Institutes of Health Research, CAN-BIND (partially funded by the Ontario Brain Institute), Fonds de recherche du Québec (FRQS), and Janssen Research & Development, LLC. Additional funding for CAN-BIND was provided by the CIHR, Brain Canada, Lundbeck, Bristol-Myers Squibb, Pfizer, and Servier. Declaration of Interest: The authors declare no competing interests Ethical Approval: Research Ethics Boards at each recruitment site approved the study design.