Autism spectrum disorder (ASD) is an early onset, developmental disorder whose genetic cause is heterogeneous and complex. In total, 70% of ASD cases are due to an unknown etiology. Among the monogenic causes of ASD, fragile X syndrome (FXS) accounts for 2-4% of ASD cases, and 60% of individuals with FXS present with ASD. Epigenetic changes, specifically DNA methylation, which modulates gene expression levels, play a significant role in the pathogenesis of both disorders. Thus, in this study, using the Human Methylation EPIC Bead Chip, we examined the global DNA methylation profiles of biological samples derived from 57 age-matched male participants (2-6 years old), including 23 subjects with ASD, 23 subjects with FXS with ASD (FXSA) and 11 typical developing (TD) children. After controlling for technical variation and white blood cell composition, using the conservatory threshold of the false discovery rate (FDR ≤ 0.05), in the three comparison groups, TD vs. AD, TD vs. FXSA and ASD vs. FXSA, we identified 156, 79 and 3100 differentially methylated sites (DMS), and 14, 13 and 263 differential methylation regions (DMRs). Interestingly, several genes differentially methylated among the three groups were among those listed in the SFARI Gene database, including the PAK2, GTF2I and FOXP1 genes important for brain development. Further, enrichment analyses identified pathways involved in several functions, including synaptic plasticity. Our preliminary study identified a significant role of altered DNA methylation in the pathology of ASD and FXS, suggesting that the characterization of a DNA methylation signature may help to unravel the pathogenicity of FXS and ASD and may help the development of an improved diagnostic classification of children with ASD and FXSA. In addition, it may pave the way for developing therapeutic interventions that could reverse the altered methylome profile in children with neurodevelopmental disorders.

Transcriptome-wide association analysis is a powerful approach to studying the genetic architecture of complex traits. A key component of this approach is to build a model to impute gene expression levels from genotypes by using samples with matched genotypes and gene expression data in a given tissue. However, it is challenging to develop robust and accurate imputation models with a limited sample size for any single tissue. Here, we first introduce a multi-task learning method to jointly impute gene expression in 44 human tissues. Compared with single-tissue methods, our approach achieved an average of 39% improvement in imputation accuracy and generated effective imputation models for an average of 120% more genes. We describe a summary-statistic-based testing framework that combines multiple single-tissue associations into a powerful metric to quantify the overall gene-trait association. We applied our method, called UTMOST (unified test for molecular signatures), to multiple genome-wide-association results and demonstrate its advantages over single-tissue strategies.