Decoding variants of uncertain significance in systemic autoinflammatory diseases

The development and implementation of genetic testing has revolutionized the diagnostic landscape of autoinflammatory diseases, leading to an exponential increase in the identification of disease-associated genetic variants. Yet a substantial proportion of these are considered variants of uncertain significance (VUS), complicating both diagnosis and therapeutic decision-making. This challenge is relevant not only for monogenic systemic autoinflammatory diseases, but also in the context of genetically complex disorders that can involve multiple low-penetrance variants. Advances in protein structure prediction tools, machine learning and artificial intelligence provide powerful computational frameworks for the classification of variants; however, their predictive accuracy must be benchmarked against functional assays, particularly with respect to gain-of-function variants. Functional screening approaches benefit from both technological progress and expanding knowledge of the innate immune pathways underlying systemic autoinflammatory diseases. Large-scale analyses of variants including multiplexed functional assays and deep mutational scanning experiments have enabled the assessment of hundreds of variants, notably in NLRP3, MEFV and ADA2, generating datasets that improve variant interpretation and genetic diagnosis. Altogether, these advances increase the potential of accurately predicting in the near future the effects of missense VUS, although numerous challenges remain to be addressed, especially those concerning our understanding of the influence of non-coding VUS in systemic autoinflammatory diseases.