Human leukocyte antigen (HLA) mismatching and minimization of immunosuppression are two major risk factors for the development of de novo donor-specific antibodies, which are associated with reduced kidney graft survival. Antibodies do not recognize whole HLA antigens but rather individual epitopes, which are short sequences of amino acids in accessible positions. However, compatibility is still assessed by the simple count of mismatched HLA antigens. We hypothesized that the number of mismatched epitopes, or ("epitope load") would identify patients at the highest risk of developing donor specific antibodies following minimization of immunosuppression. We determined epitope load in 89 clinical trial participants who converted from cyclosporine to everolimus 3 months after kidney transplantation. Twenty-nine participants (32.6%) developed de novo donor specific antibodies. Compared to the number of HLA mismatches, epitope load was more strongly associated with the development of donor specific antibodies. Participants with an epitope load greater than 27 had a 12-fold relative risk of developing donor-specific antibodies compared to those with an epitope load below that threshold. Using that threshold, epitope load would have missed only one participant who subsequently developed donor specific antibodies, compared to 8 missed cases based on a 6-antigen mismatch. DQ7 was the most frequent antigenic target of donor specific antibodies in our population, and some DQ7 epitopes appeared to be more frequently involved than others. Assessing epitope load before minimizing immunosuppression may be a more efficient tool to identify patients at the highest risk of allosensitization.