Why Epigenetics?

    In today’s post-genomic era, we have an abundance of sequenced genomes that have contributed greatly to our knowledge of biology and human health. However, despite this wealth of big data, we are still left with fundamental questions about the heritability of complex traits and disorders. It is becoming increasingly apparent that many complex behaviours and neurodevelopmental disorders cannot be explained by genetic sequence alone. Rather, the field of genetics is undergoing a paradigm shift, which is resulting in a synthesis of genetics with epigenetics. Epigenetic marks maintain gene expression profiles related to development and tissue specificity without altering the underlying DNA sequence. Epigenetic marks are heritable in dividing cells and yet distinct from transcription factors, however, they influence each other. There are a number of epigenetic marks, which include DNA methylation, histone post-translational modifications, and non-coding RNAs.

Through a collaboration with Canadian Epigenetics, Environment and Health Research Consortium (CEEHRC) and EpiGenie, we created the above video along with other interviews as a science communication effort.

How Does Epigenetic Variation Impact Neurodevelopment?

    The epigenome functions as a molecular interface between the environment and genome that shapes cellular development. My research examines impacts on the epigenome from both the environmental and genetic perspectives. Since we often learn about basic biological mechanisms through the examination of disorders that disrupt these mechanisms, my research is focused on the epigenomics of neurodevelopmental disorders (NDDs). By examining the most common genetic and environmental causes of NDDs, I have begun to establish a model framework that can be applied to understand the complex gene x environment interactions that are typical of other NDDs.

Genetic Impacts on the Epigenome: Down Syndrome

    Down syndrome (DS) is the most common chromosomal aneuploidy and the leading genetic cause of intellectual disability. Patients with DS also display a distinct cancer risk profile, where they have an elevated risk for leukemias but a decreased risk for solid tumors. Mechanistically, DS represents an example of genetics shaping epigenetics, as reduced representation methods have demonstrated genome-wide differences in DNA methylation in multiple tissues.
     In our research, we were the first to apply low-pass whole genome bisulfite sequencing (WGBS) to neonatal dried blood spots and also post-mortem DS brain samples. By assaying regions of the genome that have never been examined before in the brains and neonatal blood of DS patients, we confirmed and expanded on many past findings.  Leveraging over a dozen DS datasets and 127 reference epigenomes, our analyses characterized a dichotomy in the DNA methylation profiles, where the hypermethylated regions contain a pan-tissue signature and hypomethylated regions contain a tissue-specific signature. In addition to providing a general framework for epigenome-wide association studies (EWAS) that utilizes low-pass WGBS, our approaches provided fundamental biological insight into both the early and late life impacts of a chromosomal aneuploidy. 

RUNX1 DNA Methylation profile in Down Syndrome
A large block of hypermethylation in the RUNX1 super enhancer is associated with hypomethylation of RUNX1 binding sites. Mechanistically, the RUNX1 transcription factor has been previously shown to demethylate the sites it binds to.

Biotechnology and Bioinformatics: Low-Pass Whole Genome Bisulfite Sequencing 

     I have contributed to a wide range of experiments that utilized low-pass whole genome bisulfite sequencing (WGBS) to examine DNA methylation. In order to examine these diverse DNA methylation profiles using WGBS, I developed two bioinformatic repositories that are optimized for high performance computing. The first repository is CpG_Me, a Unix shell based pipeline for the pre-processing, alignment, and quality control and quality assurance (QC/QA) of WGBS data. The second repository is DMRichR, an R package and accompanying executable R script for the statistical testing and downstream analysis of differentially methylated regions (DMRs) from the aligned WGBS data. 
      My bioinformatic approaches have led to my involvement in diverse projects related to examining cell-free fetal DNA, cell culture, post-mortem brain tissue, neonatal dried blood spots, placenta, and cord blood. These projects have spanned multiple organisms, including humans and rodents. For example, we have examined the effects of augmented maternal care in a rodent model and we were the first to provide an integrative profile of the protective genome-wide differences in the DNA methylation that confer resilience to stress in the brain. We have also examined post-mortem brain tissue from patients with Rett syndrome, Dup15q Syndrome, and Idiopathic Autism Spectrum Disorders, where we uncovered that DNA methylation profiles of these neurodevelopmental disorders converge at loci related to neural-immune risk factors. Finally, we have also examined classical disorders of genomic imprinting, where we characterized the deregulation of genomically imprinted gene networks in a rodent model of Angelman Syndrome, as well as a rodent model of Prader-Willi Syndrome with a focus on circadian rhythm.

Epigenetics cover featuring circos plot
A Circos plot I generated was featured as cover art for the journal Epigenetics.

Environmental Impacts on the Epigenome: Prenatal Alcohol Exposure

     The consumption of alcohol during any stage of pregnancy can result in FASD, with their being no empirically determined safe dose or safe timing of PAE. FASD is the leading preventable cause of intellectual disability in the western world. In Canada, from 2003-2010, ~10% of pregnant women and ~20% of breast- feeding women consumed alcohol. While no official record exists for Canada, FASD is typically diagnosed in only 2-5% of children. Accurate biomarkers of PAE and its neurodevelopmental impacts are needed, as current diagnostics are not specific to the large number of individuals in the spectrum without craniofacial abnormalities. During my graduate research, we utilized a mouse model of a chronic but moderate prenatal alcohol exposure throughout pregnancy and integrated the genome-wide DNA methylation, microRNA expression, and gene expression profiles of adult mouse brains. Strikingly, while there were genome-wide alterations at all levels, there was a distinct profile of alterations to genomically imprinted clusters of non-coding RNA, which are expressed in a parent-of-origin specific manner and are involved in (neuro)development. 
    After establishing that there are long-term alterations to the epigenome in the brains of prenatal alcohol exposed mice, we examined the DNA methylation profile in the buccal swabs of young children with FASD and compared the results to our mouse model. Strikingly, we observed similar alterations to epigenetically sensitive regions that we identified in the rodent models. The results implicated conserved large differentially methylated regions (DMRs) within the clustered protocadherins, which establish single-cell neuronal identity, in both adult mouse brains and the buccal swabs of children.

Clustered protocahderin DMR in human children with FASD and prenatal alcohol exposed mice
The hypermethylated clustered protocadherin DMRs that are conserved between the buccal swabs of human children with FASD and the adult brains of a mouse model of PAE.

Deconvoluting the Epigenomics of Psychosis ​

     I collaborated with other members of my graduate lab to investigate the epigenomics of psychosis, specifically schizophrenia. This first involved collaborating with a neuroscience lab, where we examined a rat model of anti-psychotic treatment and uncovered that many of reported DNA methylation differences in patients with schizophrenia appear to be due to their antipsychotic treatment rather than the disorder itself. Second, using this knowledge as a translational map, we examined the genome and epigenome of monozygotic twins discordant for schizophrenia and uncovered a number of similar and unique alterations. Together, these findings are cited by researchers examining a variety of psychiatric disorders as words of caution for interpreting epigenomic associations that are confounded by medication exposure.