oa Inheritance Of Methylation In The Qatari Population

Background & Objectives: The environment surrounding an organism can have a direct impact on methylation and trans-generational effects in both mammals [1] and plants [2, 3]. Some epigenetic changes are heritable or somehow passed on to future generations without altering the organism's underlying DNA sequence[4]. The concept of DNA methylation heritability in mammals is emerging, although the mechanism through which environmental impact is epigenetically transmitted is still not clearly understood. We search for methylation sites that exist in three states (trimodal) on a population level in a Qatari population for the first time. Using this subset of sites, we can test for a pattern of Mendelian inheritance of methylation, specifically by looking at trios. These patterns are then compared to randomized individuals where the trio structure was disrupted to conclude on the effect of family structure towards DNA methylation heritability. Methods: We used a Qatari cohort consisting of families. Some of the families contained trios (father, mother, and offspring). We obtained both methylation data (Illumina 450K array) [5] and whole genome sequencing data (Illumina Hiseq2500 platform) for the samples. We tested the 450K methylation sites for trimodality and filtered those satisfying the definition. The trimodal sites were then tested for Mendelian agreements and violations only in the trios. This was compared to randomized samples having no parent-offspring relationship as an enrichment study. Finally, we checked the trimodal sites for the likely possibility of being explained by an underlying genetic association. Results: We found less than 1% of the methylation sites across the dataset to meet the predefined trimodal criterion. We report that non-related individuals have a higher average of trimodal sites showing Mendelian violation in comparison with actual trios. Moreover, the average of trimodal sites exhibiting Mendelian agreement in trios is significantly higher in comparison with randomized samples. We also studied the effect of cis-acting SNPs on explaining the observed patterns of methylation heritability. Although in the majority of the sites, the pattern is elucidated by SNPs, there still a minute number of sites where the methylation is not explained by SNPs. These findings are replicated and validated in a bigger cohort. Conclusions: The great majority of methylation sites that follow a Mendelian inheritance pattern in trios are driven by an underlying SNP association. However, a minor percent of the trimodal sites we studied did not have a cis-acting SNP association although this does not rule out the effect of trans-acting SNP associations. We can speculate on other factors that could be causative of the trimodal pattern such as correlations between neighboring CpGs, SNPs otherwise in the genome, or even genome-wide probe-SNP eQTL associations.