Epigenome analysis of autism spectrum disorder samples

The aim of this study was to identify ASD vulnerability components, explore possible vulnerability subgroups, and evaluate environmental contribution to phenotypic variability. We performed PCA and cluster analysis using ASD risk factor data. We found that PC1 had a higher correlation with psychosocial stress (maternal stress, maternal education, and social class), and PC2 a higher correlation with biological factors (Psychiatric Family History and Gestational Complications). Clustering analysis of PCA scores using bootstrap approach showed two clusters. Comparing the methylome between clusters we found 11.879 DMPs (p<0.05), and their CpG sites were enriched in VMRs, most showing environmental and genetic influences. Hypermethylated probes presented higher rates in different regulatory regions associated with functional SNPs, indicating that the subgroups may have different affected regulatory regions and their liability to disease explained by common variations. Vulnerability score moderated by epigenetic clock was associated with Vineland Total score (p=0.0036, adjR2=0.31), suggesting risk factors with stress burden can influence phenotype. DNA from whole blood samples were treated using the EZ DNA Methylation kit (Zymo Research, Irvine, California,USA). The bisulfite converted DNA was hybridized in the BeadChip Human Methylation 450 microarrays (Illumina Inc., San Diego, California, USA) following the manufacturer’s protocol. Raw data was extracted by iScan SQ scanner using the GenomeStudio software (v.2011.1), with the methylation module v.1.9.0 IDAT files. Overall there 66 samples of ASD with diagnosed using ADI-R. IQ assessed by the SON-R 2½-7. Their functionality was assessed with Vineland, ASD severity with the CARS, and the patients’ internalizing and externalizing symptoms with the CBCL.