A novel clustering approach for the analysis of longitudinal islet autoantibody profiles

Background/Aim: Progression to type 1 diabetes (T1D) is not uniform. Based on individual genetic background and environment children develop islet autoimmunity at different ages, show different autoantibody profiles, and progress to T1D rapidly, slowly, or not at all. Here, we aimed to develop a mathematical algorithm to model complex longitudinal autoantibody profiles, and identify clusters that are associated with different progression of T1D. Methods: We analysed data from 156 children who have been followed from birth in the BABYDIAB study and developed persistent autoantibodies against insulin (IAA), glutam…