Marie Camerota My name’s Marie Camerota, and  I’m an Assistant Professor in the Department of   Psychiatry and Human Behavior, at the Warren  Alpert Medical School of Brown University.   I’m located in Providence, Rhode Island,  in the United States. I’m a Developmental   Psychologist, and my research focuses on  neurodevelopment in high risk paediatric   populations. Today, I’m going to be  talking about epigenetics and how   epigenetic research is being used to advance  our understanding of child neurodevelopment. The objectives of my talk today are to provide an  overview of epigenetic concepts and mechanisms,   to summarise the history of behavioural  epigenetic research, to describe methods   for studying epigenetics in human research, to  highlight current understanding of the role of   epigenetics in neurodevelopment, and we’ll  end with a discussion of future directions. So, to begin to understand the field of  epigenetics, we must first have a good   understanding of genetics, so let’s review. Each  cell in our body contains 23 pairs of chromosomes.   Chromosomes contain tightly coiled strands  of DNA that are organised into genes. Now,   the genetic code refers to the instructions  that determine what protein a gene will make.   The genetic code is made up of four chemical  bases adenine, guanine, cytosine and thymine.   And we can see that DNA looks a bit like  a stepladder, and this is because the four   bases pair with one other, where adenine pairs  with thymine, and cytosine pairs with guanine,   and together, these base pairs form  all of the rungs of the DNA stepladder. Our genetic code is expressed via two  processes called transcription and   translation. In transcription,  double helix DNA is unzipped and   RNA polymerase synthesises messenger  RNA using complementary base pairing,   essentially, this makes a copy of DNA. And  the reason we do this is to protect DNA,   we want to make a copy of it, rather than  altering or acting upon the underlying sequence. After mRNA is created, it moves to the cytoplasm,  and a ribosome then uses mRNA as a template, to   assemble a specific protein. Each group of three  bases in the RNA codes for a specific amino acid,   which is added to the growing protein chain.  The protein will then go off to do its job   in the body, whether it’s a hormone, an antibody,  an enzyme or something with many other functions. When we talk about gene expression, essentially,  what we mean is that we are making a gene’s   protein. Now, when we talk about epigenetics, epi  literally means on top of, or, in addition to,   so epigenetics describes processes and  mechanisms that operate on top of DNA,   and can impact gene activity  without changing the genetic code. So, let’s think about epigenetic marks  using a musical analogy. Imagine you’re   sitting listening to a piece of music,  the musical notes are like the DNA,   they are the instructions that tell musicians what  notes to play. We can think about epigenetics as   musical notations. These are the markings  that you add on top of the musical notes,   that tell you how to play the music, whether it’s  quieter or louder or in a crescendo. Ultimately,   the combination of genetics and epigenetics,  or the musical notes and the markings,   determine the phenotype, or what  music you actually end up hearing. Epigenetics can also help us understand  why individuals with the same genotype   can have different phenotypes. So,  for example, even identical twins,   who share 100% of their DNA, do not look  100% identical, especially as times goes   on. They also do not have the exact  same risk of different diseases. And   this is because the environment can change gene  expression, therefore, epigenetics also refers   to how an individual’s behaviours and their  environment can change their gene expression. Let’s look at an example. So, it’s hard to  believe, but these two mice are actually   genetically identical. They have the same form of  the agouti gene that controls coat colour, yet,   one mouse on the left is large and yellow, and  the other is small and brown, and the reason for   this is that they have differential expression of  the same gene. When the agouti gene is expressed,   it leads to yellow coat colour, when  the agouti gene is not expressed,   the more brown the coat colour becomes. This  gene also impacts metabolism, and so the yellow   mouse also has a higher risk of obesity and  diabetes, compared to the small, brown mouse. And one of the factors that have – has been  found to impact the expression of the agouti   gene is maternal diet. So, when mice mothers  were fed a diet high in certain nutrients,   they gave birth to mice who had  browner coats, and, therefore,   this explains how the environment, in this case  prenatal nutrition, can impact a gene expression. So, how do genes become more or less expressed?  Epigenetic processes can alter gene expression by   interfering at two critical stages of gene  expression that we reviewed earlier. They can   either block DNA from being transcribed into RNA,  so they can interfere at that first step, or they   can stop RNA being translated into proteins, so  they can stop gene expression at that second step. The epigenetic mechanisms that we’re familiar  with that block transcription are DNA methylation,   histone modification, and chromatin remodelling,  and we’ll talk about each of these in more detail.   We know less about the epigenetic processes that  block translation, but they include the effect   of microRNAs, which can bind to messenger  RNA, to prevent it from being translated. Let’s take a closer look at the  epigenetic mechanisms that block   transcription. We have about six feet  of DNA within each cell of our body,   so we can understand that to fit into the  cell nucleus, the DNA has to be tightly   compacted. The way this happens is that DNA  is wound around proteins called histones,   and these form the bead on a string structure  called nucleosomes. Nucleosomes organise   themselves into condensed material called  chromatin, which make up the chromosome. Histone modification describes chemical  modifications of the histone tails,   via the addition of small molecules, that  can impact whether that region of DNA is   transcribed or not. And histone modification can  block transcription chemically, by not allowing   transcription factors to bind, or physically,  by determining how tightly condensed DNA is. Similarly, chromatin remodelling  can either be tight or open. So,   here, we see an example where the chromatin  packaging is tight, the DNA is not accessible,   and translic – transcription cannot occur; it  is physically blocked. When the chromatin has an   open structure, like this, we can see the  DNA is accessible, and it can be transcribed. The third, and most commonly studied epigenetic  mechanism, is DNA methylation. If you remember   from a few slides back, cytosine is one of the  four base pairs that makes up DNA. DNA methylation   describes the addition of a methyl group, or this  CH3 molecule, to the fifth carbon of the cytosine,   to create what’s called 5-methylcytosine. And even  though it’s a very small molecule, the addition of   the methyl group changes the chemical structure  of DNA and has implications for transcription. Here we see a small snippet of DNA. Methylation  typically occurs in regions of DNA where a   cytosine nucleotide is followed by a guanine  nucleotide, linked by a phosphate group. This   is termed a CpG site. In this example, the two  CpG sites are unmethylated. When a methyl group   is added to the cytosine base, we now have a  methylated CpG site. Because the methyl group   is added on top of the DNA, it does not change the  underlying genetic code or the sequence of Cs, Gs,   Ts and As, but it can change the likelihood of  that portion of the DNA sequence being expressed. Let’s take a closer look at how that happens.  Here’s another snippet of a gene. The promoter   is the region of a gene generally upstream of a  gene’s coding region, where transcription factors   and RNA polymerase bind to initiate transcription  of that gene. In this image, the CpG sites and   the promoter are primarily unmethylated,  so these are the white circles. Generally,   a lack of methylation in the promoter region means  that transcription factors can bind, transcription   can begin, and the gene will be expressed. On the other hand, when there’s a high level   of methylation in the promoter region, here  we see the black circles, it can inhibit   transcription factors from binding, which  means transcription cannot begin and the gene   will not be expressed. Now, we often talk about  epigenetic processes, turning genes on or off,   but it’s not always an all or nothing effect.  Often, DNA methylation operates like a dimmer   switch, that can reduce or enhance the amount  of gene expression, and by changing the rate of   gene expression, DNA methylation can have wide  reaching consequences for human functioning. Epigenetics is a part of normal development  and, in fact, the term was originally used   to describe the process of normal embryonic  development. Immediately following fertilisation,   cells are undifferentiated, which means they’re  not specialised and they are unmethylated. During   embryogenesis, cells become specialised  and this happens via DNA methylation. So, for example, in cells  that are becoming neurons,   the genes for making blood cells have  to be turned off or methylated, whereas,   the genes for making neurons have to  remain on or unmethylated. And this   leads to the derivation of specialised cells  like neurons, blood cells, immune cells, etc.,   each of which has its own specialised function  and its own unique pattern of DNA methylation. Now, embryogenesis is not the only period when  DNA methylation is involved in development. The   epigenome changes across development, and some  research shows that its most dynamic during the   earliest years of life. We know that both genetics  and environment can impact methylation patterns,   so which version of a gene you have  could impact the amount of methylation   of that gene. And we already talked  about the environment as an example,   in the mouse model, and we’ll talk  much more about this later, as well. Epigenetic changes are reversible, so they can  come and go across development. And, finally,   though it’s still controversial in human studies,  epigenetic changes might be able to be passed down   from parents to children and, thus, this could be  another method of intergenerational inheritance,   in addition to genetics. Again though, this  remains to be determined conclusively in humans. Interest in epigenetic research has greatly  increased in the past 20 years. The figure   here shows the number of papers published  per year since the year 2000 with the keyword   “epigenetics,” those are the blue bars, and  with the search terms “human epigenetics,”   shown in yellow. Just to give some historical  perspective, 1942 was the year when the term   “epigenetics” was first coined, and in 1983, there  was a discovery that altered DNA methylation was   found in some cancers, leading to a dramatic  increase in research on epigenetics and cancer,   and, in fact, the field of epigenetics is still  dominated by cancer research. In the year 2006,   the first epigenetic drugs were approved by  the FDA to treat cancer, and around 2014,   the FDA approved the first  cancer screening tools that   use epigenetics as a biomarker to detect  cancer-specific DNA methylation patterns. In contrast to epigenetic studies in  cancer research, there’s a shorter   history of research on epigenetics as they  relate to human behaviour and development,   and we’ll turn next to reviewing some of  those critical early behavioural studies. Michael Meany is a Scientist who’s most well-known  for his research on maternal caregiving,   gene expression and stress, and most of his work  was done in animal models. So, let’s quickly talk   about what makes a good or bad rat mother. A  good rat mother spends a lot of time licking   and grooming her pups, which provides a lot of  touch and a lot of physical stimulation. She also   arches her back to make ample room for rat pups  to nurse. A bad rat mother provides low levels   of licking and grooming, and less arched back  behaviour. And we’ll refer to these good moms as   the “high licking and grooming moms,” and the less  good moms as the “low licking and grooming moms.” Meaney first discovered that the quality  of maternal caregiving was associated   with stress reactivity in rat pups, with  pups who were exposed to high licking and   grooming having lower stress reactivity, as  compared to pups exposed to low licking and   grooming who had higher stress reactivity. Let’s take a closer look. This was studied   by exposing rat pups to a stressor, such as,  being removed from their mother and handled   by a Researcher, and observing how their stress  hormones changed. This is shown in the graph on   the left. So, a critical observation was  that the high licking and grooming pups,   these are the ones shown in the white dots,  showed less of an increase in stress hormone   following the stressor, compared to pups with  low licking and grooming mothers. These are   shown with the black dots. And to highlight the  specific importance of the caregiving environment,   rather than genetics, follow-up studies showed  that the effects of maternal licking and   grooming can be reversed by adopting pups out  to mothers with a different caregiving style. So, we know that rat pups with high  licking and grooming mothers tend to   show higher licking and grooming behaviour  themselves when they have children, whereas,   rat pups with low licking and grooming mothers  also tend to show less licking and grooming   behaviour themselves with their own offspring.  However, when rat pups born to high licking and   grooming mothers were raised or adopted out to low  licking and grooming mothers, they later showed   licking and grooming behaviour more similar to  rat pups born to low licking and grooming mothers.   Whereas, rat pups born to low licking and  grooming mothers, but raised by high licking   and grooming mothers, had high licking and  grooming behaviour themselves, and it was more   similar to rat pups who were born to, and stayed  with, their high licking and grooming mothers. So, these experiments showed that it is  the caregiving environment, specifically,   that programmes stress response and  caregiving behaviour in rats. But how   does this happen? To better understand the  biological underpinnings of this finding,   Researchers looked towards the NR3C1  gene, which codes for glucocorticoid   receptors. These are the receptors that  stress hormones like cortisol bind to. When individuals have more glucocorticoid  receptors, there’s less circulating cortisol and,   thus, a lower physiological response to stress.  Whereas, individuals with fewer glucocorticoid   receptors have higher circulating cortisol and,  thus, a greater physiological stress response.  In a follow-up of the licking and grooming work,   Weaver and colleagues reported that rat  pups exposed to high licking and grooming,   these are shown in the white bars, had lower  methylation of NR3C1 at several CpG sites,   compared to rat pups with low licking and grooming  moms. These are the ones shown in the black bars. Lower methylation in the high licking  and grooming pups means that they had   more NR3C1 gene expression,  more glucocorticoid receptors,   and, thus, a lower stress response.  Whereas, the opposite was true for   the low licking and grooming pups who had more  methylation, fewer glucocorticoid receptors,   and a greater stress response. This was a really  critical discovery, showing how maternal behaviour   could programme the epigenome in ways that have  relevance for human – or for behavioural outcomes. Now, in humans, we don’t lick our babies, usually,  so to understand whether similar processes   operate in humans, we have to look at different  types of behaviour. We know that variations in   maternal mood are associated with caregiving  behaviours. So, in an attempt to apply the   Meaney rat work to humans, Oberlander and  colleagues investigated the associations   between maternal depression during pregnancy,  methylation of NR3C1, and infant stress response. This study found that higher levels of depression  were associated with increased DNA methylation of   NR3C1. So, remember, more methylation means less  gene expression, fewer glucocorticoid receptors,   and we would predict a higher stress response. And  that’s exactly what they found, higher methylation   of NR3C1 in newborns was associated with a greater  cortisol stress response at three months of age. Another study looked at maternal breastfeeding  as a maternal behaviour in relation to NR3C1 and,   in a way, this is a closer replication of the  Meaney rat work, because it’s looking directly   at maternal behaviour, rather than as – looking  at maternal mood as a proxy. This study found   that exposure to higher levels of breastfeeding,  so the high breastfeeding group is shown in red,   was associated with decreased DNA methylation of  NR3C1, compared to infants who received low levels   of breastfeeding. These are shown in the blue. So, in turn, higher levels of methylation, again,   remember more methylation means less gene  expression, fewer glucocorticoid receptors,   was associated with an increased  stress response in infants. So,   together it appears that what was true in  mice also appears to be true in humans. NR3C1 has definitely been the most studied  gene in methylation studies among humans.   And what we have learned so far is that  a number of environmental experiences are   associated with its methylation.  So, factors like prenatal stress,   trauma exposure, other forms of early life  adversity, are all positively associated,   so they predict more methylation of this gene.  Whereas, positive factors in the environment,   like, maternal sensitivity, are associated with  less methylation. And then, in turn, greater   methylation of NR3C1 is associated with atypical  newborn neurobehaviour, increased behaviour   problems and symptoms of psychopathology,  and poor socioemotional functioning. So, the research reviewed so far on NR3C1,  all of these are examples of candidate gene   studies. These types of studies investigate  methylation at a small number of CpG sites   within specific genes. So, for example,  studies of NR3C1 all tend to focus on CpG   sites in a specific gene region, which  is highlighted in grey in this figure,   this is the promoter region. And then  under the grey box, we see the specific   sequence of genetic code for that region, with  the CpG sites highlighted in red. And this allows   Researchers to take a very detailed look into the  regulation of specific genes, and to be consistent   with one another in terms of which regions  within a gene are specifically of interest. Candidate gene studies also pose specific  hypotheses about how methylation should be   associated with environmental exposures  or with outcomes. So, for example,   if we were doing a new NR3C1 study, we  might hypothesise that higher levels of   methylation of NR3C1 would be associated with  worse behavioural outcomes for children. But   we know that humans have about 20,000 genes,  and that these genes work together to produce   our physical and biological traits, therefore,  it’s somewhat naive to think that we can study   individual genes to understand something  as complex as human health or behaviour. Thus, while candidate gene studies  were historically very popular,   the fields of epigenetics is now moving towards  epigenome-wide association studies, or EWAS,   as a way of uncovering novel genes that may  be related to human traits. EWAS allow us   to study a large number of CpGs throughout the  entire genome, and these types of studies tend   to be more hypothesis free, so these are more  focused on the discovery of novel CpG sites. EWAS take advantage of the fact that  humans differ in their DNA methylation   levels at many CpGs within the genome, and  by testing associations between methylation   levels and outcomes, or predictors,  EWAS attempt to identify the CpGs   that are the most strongly associated with  environmental exposures, or with human traits. This figure is called a Manhattan plot, and each  dot in the picture represents one CpG site. The   X axis shows the chromosome location of the  CpG site and the Y axis shows the Log P value,   and we can interpret this as showing the strength  of the association of that CpG with the variable   of interest. In this case, we studied DNA  methylation of about 450,000 CpGs in newborns,   and related their methylation level to  children’s later attention problems. In an EWAS, because you’re conducting so many  statistical tests, there’s a need to have a   stricter cutoff for statistical significance  than the typical p less than .05. So, here,   the blue line and the red line show two different  thresholds for statistical significance,   where red is the stricter threshold. So, we  see that there’s a number of CpGs above the   more lenient threshold, and three CpGs  that are above the stricter threshold,   and these are the ones we might focus on as being  the most predictive of later attention problems. Now, sometimes, the results from an EWAS  could be used to identify important genes,   which could then be studied in more  detail via candidate gene work. So,   these two methods could tie into one another  and be used in a complementary way. Now,   both candidate gene and EWAS approaches have  been applied to studying the links between   epigenetics and neurodevelopment in children.  And when we talk about neurodevelopment,   you might first think that we’re talking about  the brain, differences in brain structure or   differences in brain function, and that is  certainly one aspect of neurodevelopment.   But we’re, also, more often, talking about  the development in skills like attention,   executive function, language, reasoning, memory,  in addition to things like behaviour problems,   and symptoms of mental health conditions.  And there’s relatively more research on   epigenetics and these neurodevelopmental domains,  as opposed to epigenetics and brain development. The first overarching theme from this body of  research is that many of the biological and   environmental factors that we know to be  associated with neurodevelopment are also   associated with differences in DNA methylation.  This includes pre and perinatal risk factors,   like birth weight, gestational age, and  maternal smoking during pregnancy, as well as,   prenatal maternal stress, maternal mood disorders,  like depression, and prenatal malnutrition. And,   importantly, all of these risk factors have  been studied in multiple cohorts of children,   and together these studies lend support for  prenatal programming hypotheses that state   that maternal experiences and health during her  pregnancy can have long lasting implications   for children’s health and development, via  their effect on children’s DNA methylation. Chemical exposures, like air  pollution, and heavy metals,   have also been linked to differences  in methylation in children, therefore,   DNA methylation might explain how these chemicals  have an impact on children’s development. There’s   some positive news too. Earlier we talked  about the links between maternal caregiving,   both in mice and in humans, and methylation  of the NR3C1 gene. In this recent EWAS,   Researchers investigated associations between  maternal sensitive caregiving, when children   were ages three and four, and DNA methylation  from children’s blood, collected at age six. And they found that there were four gene  regions where DNA methylation was robustly   associated with maternal sensitivity, and  these regions have previously been implicated   in psychological and developmental problems,  inflammation, and stress response. And so,   these four gene regions may potentially be  important candidate genes to study further. So,   if we look at all the evidence together, it  seems to suggest that one of the ways that   the prenatal and postnatal environment  can alter children’s neurodevelopmental   trajectories is through their impact on  children’s DNA methylation, and epigenetics   may potentially be a mechanism that explains  the biological embedding of early experience.  The second big thing we know is that we  can measure differences in children’s DNA   methylation patterns early in life, and  that these patterns are related to their   neurodevelopmental outcomes in later childhood.  Attention-deficit hyperactivity disorder, or ADHD,   and autism spectrum disorder, or ASD, are two of  the most common and well-known neurodevelopmental   disorders. ADHD has a prevalence of eight  to 11%, whereas ASD is closer to 1 to 3%. We know that there is a genetic basis to both  of these disorders, and recent studies show   that differences in DNA methylation are also  associated with ADHD and ASD risk. In addition,   histone modifications have been shown  to be different in individuals with ASD,   versus those without. Honing in on ADHD  specifically, this neurodevelopmental   outcome has been studied in at least eight EWAS  to date. Multiple studies have found methylation   of the VIPR2 gene to be related to childhood  ADHD diagnosis, and the TP73 gene has been   related to both childhood ADHD diagnosis,  and to attention problems in toddlers. VIPR2 is a gene that codes for  a neuropeptide, that works as a   neurotransmitter and neuroendocrine hormone,  and regulates processes related to mood and   behaviour, including circadian rhythm.  Whereas, the TP73 gene encodes for a   transcription factor involved in cellular  response to development and stress, yet,   its role in early development is still unclear.  So, together, the results from these studies   are potentially identifying promising  candidate genes worthy of future study. Methylation of the HES1 gene has also been  validated in multiple cohorts, in terms of   its relation to child neurodevelopment. Higher  methylation of HES1 was shown to be related to   higher IQ in children at age four, and in  a separate group of children, increased   methylation of HES1 was shown to be associated  with higher executive function. And the HES1   gene has been shown to be involved in cellular  differentiation of the central nervous system. So, it’s worth noting at this point that all  the effect sizes for the studies mentioned so   far have been relatively small and, thus, it’s  unlikely that methylation of any of these single   genes alone can predict neurodevelopmental  outcomes. And we are still a far ways away   from using methylation patterns to either screen  for, or to treat, neurodevelopmental disorders. We have also begun to study how differences in  DNA methylation at birth can help predict outcomes   for children at high risk for neurodevelopmental  impairment. So, in work by myself and our team,   we’ve studied DNA methylation in children  born very preterm. These are children born   less than 30 weeks gestational age. And,  so far, we’ve demonstrated associations   between DNA methylation at birth and neonatal  neurobehaviour, attention problems at age two,   and cognitive ability at age three. And these  studies show how epigenetics could potentially   be used to help identify high risk infants who  have a higher chance of developing long-term   neurodevelopmental problems. Though, again, the  effect sizes associated with single CpG sites   tend to be small, and we’re not yet at the point  of developing diagnostic tools for these infants. Another high risk population that’s  been studied from an epigenetic lens   are infants exposed prenatally to opioids.  We know that the opioid receptor gene,   OPRM1, is associated with vulnerability to opioid  addiction, and that differences in methylation of   this gene are associated with the severity of  neonatal opioid withdrawal among children who   are exposed to opioids during pregnancy.  Neonatal opioid withdrawal is typically   treated pharmacologically by providing  the newborn with morphine after birth. In this study, we demonstrated that there is a  significant decrease in methylation of OPRM1,   following pharmacological treatment  for NAS, and that infants with greater   decreases in methylation of OPRM1 show more  improvements in their neurobehaviour. So,   this research shows that epigenetics can  potentially be used to understand biological   mechanisms underlying treatment response  for certain conditions impacting children. Thus, we now have convincing evidence that  biological and environmental risk factors   that impact neurodevelopment can also have  an impact on children’s DNA methylation,   and that differences in children’s DNA methylation  are associated with neurodevelopmental outcomes,   including risk for neurodevelopmental  disorders like ADHD. And we’re even   beginning to study epigenetic patterns as  a mechanism that may explain the impact   of environmental and biological risk  factors on neurodevelopmental outcomes. But there’s still a lot left to learn, so where  do we go next? Most of the studies of epigenetics   in humans use peripheral tissue. So these are  samples that are collected from cheek swabs,   saliva, or blood samples, and they use these  samples to measure DNA methylation patterns.   Earlier, I mentioned that different cells  each have different methylation patterns,   that’s what makes a brain cell different from  a blood cell, for example. So, even if we find   associations between DNA methylation patterns in  peripheral tissue and neurodevelopmental outcomes,   it’s unclear what these methylation patterns are  telling us about processes occurring in the brain. In adults, there are databases that can be used  to compare methylation patterns in peripheral   tissue versus brain tissue, but, as of now, these  types of resources are not available for child   studies. Therefore, it’s still unclear whether  the results from peripheral tissue are telling   us anything mechanistic about neurodevelopment.  That is, do neurode – or do methylation patterns   cause differences in neurodevelopmental  outcomes, or are they merely correlated? Now it’s important to note that DNA methylation  could be a biomarker of neurodevelopmental   outcomes without being causally related,  and this could happen if DNA methylation   patterns reflect differences in exposure to  risk factors, disease onset, progression,   treatment response, etc., but do not themselves  cause neurodevelopment, all that matters is that   the methylation patterns are correlated with  the outcome. And if we’re thinking about the   promise of epigenetics as a biomarker, the type  of tissue that it’s collected from matters less,   and, in fact, peripheral tissue would  be preferred for these types of uses,   because the peripheral tissue, like  cheek swabs, are easily accessible. Another important future direction is  investigating changes in epigenetic   patterns over time. We don’t yet know the  extent to which genes become more or less   methylated over development, or how these  changes in methylation track with changes   in neurodevelopmental outcomes. Preliminary  evidence suggests that the epigenome is the   most dynamic early in life, but this  remains to be more closely studied,   to understand what the rate of change is  during different developmental periods. We also don’t yet know when the epigenome is the  most sensitive to environmental influences. Again,   early life seems to be particularly important,  but it’s possible that there are other sensitive   periods, where the environment has a  greater impact on methylation patterns,   so, perhaps, another sensitive  period could be the pubertal period. And, finally, it’s important that we begin to  integrate epigenetic data with broader genetic,   behavioural and environmental data, to better  understand for whom, and under what conditions,   epigenetic patterns relate  to outcomes. So, for example,   are these relationships stronger in children  at risk for neurodevelopmental impairment,   such as children born preterm? And by studying RNA  and proteins together with epigenetic patterns,   we can potentially learn more about the  functional impact of differences in methylation. So, in conclusion, epigenetic studies provide  an exciting opportunity to study the molecular   underpinnings of behaviour. We’re learning more  about known genes like NR3C1 and OPRM1 and, also,   discovering potentially promising  new candidate genes that may be   biomarkers for other environmental exposures  or neurodevelopmental outcomes. In the future,   this line of work could lead to the  development of epigenetically-based treatments. So, I showed the example of OPRM1 as a potential  mechanism that explains the treatment response to   morphine in opioid exposed infants. Similar  investigations could help explore whether   epigenetics can help explain child response  or non-response to stimulant medications,   for example, or to behavioural interventions.  So, for example, might the methylation of NR3C1   change in response to a stress reduction,  or a parenting skills intervention? And, finally, thinking about the use  of epigenetics as a biomarker, if we   find epigenetic signals that strongly predict  neurodevelopmental and behavioural disorders, then   we can develop novel screening tools, akin to the  cancer screening tools currently on the market,   that can identify children at risk for these  disorders earlier than is currently possible,   and this would open the door to earlier treatment  and intervention to promote positive outcomes. We’re still very much at the  beginning of this work, and,   yet, I look forward to advances to be made  in the years to come. Thank you very much,   I hope this has been an informative overview.  For those who are interested in learning more   about the fundamentals of epigenetics, I’ve  linked a few resources here. Thank you.