Deep learning algorithms reveal genomic markers for anxiety disorder in a large cohort of children with down syndrome

Despite a significant burden of neurobehavioral and psychiatric comorbidities in children with Down syndrome (DS), and the general increased risk for anxiety in individuals with intellectual disabilities (ID), children with DS have significantly lower odds of anxiety. Understanding the specific mechanisms of anxiety in DS could inform the development of new treatments. This study performed a comprehensive investigation of genomic variants that contribute to anxiety disorders in DS, as well as variants shared in other mental disorders. We employed deep learning algorithms using neural network models in combination with one of the largest whole-genome sequencing (WGS) cohorts of 1479 DS individuals and family members, including 255 DS probands diagnosed with at least one type of mental disorder, of whom 74 had confirmed anxiety disorders. We found that only a fraction (19%) of anxiety-specific corresponding gene variants previously reported overlap with those shared in anxiety in DS patients, suggesting distinct molecular mechanisms for anxiety in DS individuals. Functional overrepresentation analysis suggested that anxiety results from a complex interplay of genetic and environmental factors. Additionally, non-coding variants, particularly those proximal to splicing sites, play significant roles. Moreover, the variants associated with anxiety and other mental disorders are not uniquely distributed genome wide. Several loci, including 17q25, 16q23, 21q22, and 22q13, show greater weight in DS patients. Furthermore, 29 biomarkers containing recurrent anxiety-specific variants were identified to assist in the diagnosis of anxiety in the DS population. This pioneering study represents the first comprehensive exploration of anxiety disorders in DS utilizing WGS cohorts and advanced deep-learning AI models. The results indicate that anxiety disorder in DS patients has distinct molecular patterns from other mental disorders. The insights gained from our research offer valuable understanding of underlying mechanisms and hold promise for enhancing clinical diagnosis and potentially guiding more effective intervention strategies in this vulnerable population.