Autism spectrum disorders (ASD) and related developmental phenotypes are among the most devastating of childhood disorders in terms of lifetime challenges and costs to families. We propose to test the hypothesis that autism and related disorders have an epigenetic basis. We challenge the standard genetic paradigm for autism as incomplete and argue that environmental factors during pregnancy play a critical role in the disorder and that these are mediated by epigenetic mechanisms. We propose a substantially new approach to the etiology of autism and related disorders that integrates genetic, epigenetic, and environmental information through a series of progressive epidemiologic analyses of prospective data beginning at the onset of pregnancy, through the first years of the newborn's life. We have partnered two complementary pregnancy cohorts: 1) the Early Autism Risk Longitudinal Investigation (EARLI) Network, which is recruiting pregnant women at high risk of having a new child with ASD because they already have an autistic child, 2) the Johns Hopkins University National Children's Study site, which is recruiting representative pregnancies in two Maryland counties. Together, these cohorts an extraordinary wealth of data across pregnancy and postnatally in 800 mothers, fathers, and children including: biosamples from mothers at least twice during pregnancy, from children at birth and 12 months, from fathers during the pregnancy, and from 300 placenta; multiple pre-conception and in utero exposures documented and/or directly measured; assessment of child development features associated with ASDs measured at birth (gestational age, birth weight and head circumference) and at 12 months (language, social, cognitive skills). We will perform genome-wide methylation and allele-specific expression analyses on these biosamples to address the following questions: 1) Are there regions of the epigenome that are susceptible to environmental insults occurring before and during pregnancy?; 2) Are there regions of the epigenome that correlate with quantitative newborn and infant developmental phenotypes related to ASD?; 3) How does genetic variation influence these epigenetic findings? The major features of our approach include a) novel genome-wide epigenetic array and statistical methods, b) two complementary pregnancy cohorts, c) longitudinal epigenome analysis through pregnancy and early life, d) exposure measurements through pregnancy, e) quantitative developmental traits at birth and in early life, and f) integration of GWAS with epigenome data. This work will serve as a foundation for a new field of "Epigenetic Epidemiology" and represents an extraordinary opportunity to test the idea that genetics, epigenetics and environment interact before and through pregnancy to modulate the risk of a devastating and common disease, using a state-of-the-art epidemiological and epigenomic design. It will provide the first rigorous analysis of the relationship between nature and nurture in the human epigenome. ExpandCollapse Abstract