Generalized pustular psoriasis is a potentially life-threatening skin disease, associated with IL36RN disease alleles. IL36RN encodes the IL-36 receptor antagonist (IL-36Ra), a protein that downregulates the activity of IL-36 cytokines by blocking their receptor (IL-36R). Although generalized pustular psoriasis can be treated with IL-36R inhibitors, the structural underpinnings of the IL-36Ra/IL-36R interaction remain poorly understood. In this study, we sought to address this question by systematically investigating the effects of IL36RN sequence changes. We experimentally characterized the effects of 30 IL36RN variants on protein stability. In parallel, we used a machinelearning tool (Rhapsody) to analyze the IL-36Ra three-dimensional structure and predict the impact of all possible amino acid substitutions. This integrated approach identified 21 amino acids that are essential for IL-36Ra stability. We next investigated the effects of IL36RN changes on IL-36Ra/IL-36R binding and IL-36R signaling. Combining invitro assays and machine learning with a second program (mCSM), we identified 13 amino acids that are critical for IL-36Ra/IL36R engagement. Finally, we experimentally validated three representative predictions, further confirming the reliability of Rhapsody and mCSM. These findings shed light on the structural determinants of IL-36Ra activity, with potential to facilitate the design of new IL-36 inhibitors and aid the interpretation of IL36RN variants in diagnostic settings.