Eight advanced durum-breeding lines were treated with 5-methyl-azacytidine to test the feasibility of generating sources of Fusarium head blight (FHB) resistance. Of the 800 treated seeds, 415 germinated and were advanced up to four (M4) generations by selfing. Thirty-two of the resulting 415 M4 lines were selected following preliminary screening and were further tested for FHB resistance for three years at two field locations, and in the greenhouse. Five of the 32 M4 lines showed less than 30% disease severity, as compared to the parental lines and susceptible checks. Fusarium-damaged kernels and deoxynivalenol analyses supported the findings of the field and greenhouse disease assessments. Two of the most resistant M4 lines were crossed to a susceptible parent, advanced to third generation (BC1:F3) and were tested for stability and inheritance of the resistance. About, one third of the BC1:F3 lines showed FHB resistance similar to their M4 parents. The overall methylation levels (%) were compared using FASTmC method, which did not show a significant difference between M4 and parental lines. However, transcriptome analysis of one M4 line revealed significant number of differentially expressed genes related to biosynthesis of secondary metabolites, MAPK signaling, photosynthesis, starch and sucrose metabolism, plant hormone signal transduction and plant-pathogen interaction pathways, which may have helped in improved FHB resistance.