Fairness and bias correction in machine learning for depression prediction across four study populations.

A significant level of stigma and inequality exists in mental healthcare, especially in under-served populations. Inequalities are reflected in the data collected for scientific purposes. When not properly accounted for, machine learning (ML) models learned from data can reinforce these structural inequalities or biases. Here, we present a systematic study of bias in ML models designed to predict depression in four different case studies covering different countries and populations. We find that standard ML approaches regularly present biased behaviors. We also show that mitigation techniques, both standard and our own post-hoc method, can be effective in reducing the level of unfair bias. There is no one best ML model for depression prediction that provides equality of outcomes. This emphasizes the importance of analyzing fairness during model selection and transparent reporting about the impact of debiasing interventions. Finally, we also identify positive habits and open challenges that practitioners could follow to enhance fairness in their models.

What makes clocks tick? Characterizing developmental dynamics of adult epigenetic clock sites.

DNA methylation (DNAm) at specific sites can be used to calculate 'epigenetic clocks', which in adulthood are used as indicators of age(ing). However, little is known about how these clock sites 'behave' during development and what factors influence their variability in early life. This knowledge could be used to optimize healthy aging well before the onset of age-related conditions. Here, we leveraged results from two longitudinal population-based cohorts (N=5,019 samples from 2,348 individuals) to characterize trajectories of adult clock sites from birth to early adulthood. We find that clock sites (i) diverge widely in their developmental trajectories, often showing non-linear change over time; (ii) are substantially more likely than non-clock sites to vary between individuals already from birth, differences that are predictive of DNAm variation at later ages; and (iii) show enrichment for genetic and prenatal environmental exposures, supporting an early-origins perspective to epigenetic aging.

Early-life stress and the gut microbiome: A comprehensive population-based investigation.

Early-life stress (ELS) has been robustly associated with a range of poor mental and physical health outcomes. Recent studies implicate the gut microbiome in stress-related mental, cardio-metabolic and immune health problems, but research on humans is scarce and thus far often based on small, selected samples, often using retrospective reports of ELS. We examined associations between ELS and the human gut microbiome in a large, population-based study of children. ELS was measured prospectively from birth to 10 years of age in 2,004 children from the Generation R Study. We studied overall ELS, as well as unique effects of five different ELS domains, including life events, contextual risk, parental risk, interpersonal risk, and direct victimization. Stool microbiome was assessed using 16S rRNA sequencing at age 10 years and data were analyzed at multiple levels (i.e. α- and ß-diversity indices, individual genera and predicted functional pathways). In addition, we explored potential mediators of ELS-microbiome associations, including diet at age 8 and body mass index at 10 years. While no associations were observed between overall ELS (composite score of five domains) and the microbiome after multiple testing correction, contextual risk - a specific ELS domain related to socio-economic stress, including risk factors such as financial difficulties and low maternal education - was significantly associated with microbiome variability. This ELS domain was associated with lower α-diversity, with ß-diversity, and with predicted functional pathways involved, amongst others, in tryptophan biosynthesis. These associations were in part mediated by overall diet quality, a pro-inflammatory diet, fiber intake, and body mass index (BMI). These results suggest that stress related to socio-economic adversity - but not overall early life stress - is associated with a less diverse microbiome in the general population, and that this association may in part be explained by poorer diet and higher BMI. Future research is needed to test causality and to establish whether modifiable factors such as diet could be used to mitigate the negative effects of socio-economic adversity on the microbiome and related health consequences.

Are some children genetically predisposed to poor sleep? A polygenic risk study in the general population.

BACKGROUND: Twin studies show moderate heritability of sleep traits: 40% for insomnia symptoms and 46% for sleep duration. Genome-wide association studies (GWAS) have identified genetic variants involved in insomnia and sleep duration in adults, but it is unknown whether these variants affect sleep during early development. We assessed whether polygenic risk scores for insomnia (PRS-I) and sleep duration (PRS-SD) affect sleep throughout early childhood to adolescence. METHODS: We included 2,458 children of European ancestry (51% girls). Insomnia-related items of the Child Behavior Checklist were reported by mothers at child's age 1.5, 3, and 6 years. At 10-15 years, the Sleep Disturbance Scale for Children and actigraphy were assessed in a subsample (N = 975). Standardized PRS-I and PRS-SD (higher scores indicate genetic susceptibility for insomnia and longer sleep duration, respectively) were computed at multiple p-value thresholds based on largest GWAS to date. RESULTS: Children with higher PRS-I had more insomnia-related sleep problems between 1.5 and 15 years (BPRS-I < 0.001 = .09, 95% CI: 0.05; 0.14). PRS-SD was not associated with mother-reported sleep problems. A higher PRS-SD was in turn associated with longer actigraphically estimated sleep duration (BPRS-SD < 5e08 = .05, 95% CI: 0.001; 0.09) and more wake after sleep onset (BPRS-SD < 0.005 = .25, 95% CI: 0.04; 0.47) at 10-15 years, but these associations did not survive multiple testing correction. CONCLUSIONS: Children who are genetically predisposed to insomnia have more insomnia-like sleep problems, whereas those who are genetically predisposed to longer sleep have longer sleep duration, but are also more awake during the night in adolescence. This indicates that polygenic risk for sleep traits, based on GWAS in adults, affects sleep already in children.

Cord Blood Metabolite Profiles and Their Association with Autistic Traits in Childhood.

Autism Spectrum Disorder (ASD) is a diverse neurodevelopmental condition. Gene-environmental interactions in early stages of life might alter metabolic pathways, possibly contributing to ASD pathophysiology. Metabolomics may serve as a tool to identify underlying metabolic mechanisms contributing to ASD phenotype and could help to unravel its complex etiology. In a population-based, prospective cohort study among 783 mother-child pairs, cord blood serum concentrations of amino acids, non-esterified fatty acids, phospholipids, and carnitines were obtained using liquid chromatography coupled with tandem mass spectrometry. Autistic traits were measured at the children's ages of 6 (n = 716) and 13 (n = 648) years using the parent-reported Social Responsiveness Scale. Lower cord blood concentrations of SM.C.39.2 and NEFA16:1/16:0 were associated with higher autistic traits among 6-year-old children, adjusted for sex and age at outcome. After more stringent adjustment for confounders, no significant associations of cord blood metabolites and autistic traits at ages 6 and 13 were detected. Differences in lipid metabolism (SM and NEFA) might be involved in ASD-related pathways and are worth further investigation.

An epigenome-wide association study of child appetitive traits and DNA methylation.

The etiology of childhood appetitive traits is poorly understood. Early-life epigenetic processes may be involved in the developmental programming of appetite regulation in childhood. One such process is DNA methylation (DNAm), whereby a methyl group is added to a specific part of DNA, where a cytosine base is next to a guanine base, a CpG site. We meta-analyzed epigenome-wide association studies (EWASs) of cord blood DNAm and early-childhood appetitive traits. Data were from two independent cohorts: the Generation R Study (n = 1,086, Rotterdam, the Netherlands) and the Healthy Start study (n = 236, Colorado, USA). DNAm at autosomal methylation sites in cord blood was measured using the Illumina Infinium HumanMethylation450 BeadChip. Parents reported on their child's food responsiveness, emotional undereating, satiety responsiveness and food fussiness using the Children's Eating Behaviour Questionnaire at age 4-5 years. Multiple regression models were used to examine the association of DNAm (predictor) at the individual site- and regional-level (using DMRff) with each appetitive trait (outcome), adjusting for covariates. Bonferroni-correction was applied to adjust for multiple testing. There were no associations of DNAm and any appetitive trait when examining individual CpG-sites. However, when examining multiple CpGs jointly in so-called differentially methylated regions, we identified 45 associations of DNAm with food responsiveness, 7 associations of DNAm with emotional undereating, 13 associations of DNAm with satiety responsiveness, and 9 associations of DNAm with food fussiness. This study shows that DNAm in the newborn may partially explain variation in appetitive traits expressed in early childhood and provides preliminary support for early programming of child appetitive traits through DNAm. Investigating differential DNAm associated with appetitive traits could be an important first step in identifying biological pathways underlying the development of these behaviors.

Facing ostracism: micro-coding facial expressions in the Cyberball social exclusion paradigm.

BACKGROUND: Social exclusion is often measured with the Cyberball paradigm, a computerized ball-tossing game. Most Cyberball studies, however, used self-report questionnaires, leaving the data vulnerable to reporter bias, and associations with individual characteristics have been inconsistent. METHODS: In this large-scale observational study, we video-recorded 4,813 10-year-old children during Cyberball and developed a real-time micro-coding method measuring facial expressions of anger, sadness and contempt, in a multi-ethnic population-based sample. We estimated associations between facial expressions and self-reported negative feelings, explored associations of child characteristics such as sex and parental national origin with observed and self-reported feelings during social exclusion, and tested associations of observed and self-reported feelings during social exclusion with behavior problems at age 14. RESULTS: Facial expressions of sadness and anger were associated with self-reported negative feelings during the game, but not with such feelings after the game. Further, girls reported to have had less negative feelings during the game than boys, but no such sex-differences were found in total observed emotions. Likewise, children with parents of Moroccan origin reported less negative feelings during the game than Dutch children, but their facial expressions did not indicate that they were differently affected. Last, observed emotions related negatively to later internalizing problems, whereas self-report on negative feelings during the game related positively to later internalizing and externalizing problems. CONCLUSIONS: We show that facial expressions are associated with self-reported negative feelings during social exclusion, discuss that reporter-bias might be minimized using facial expressions, and find divergent associations of observed facial expressions and self-reported negative feelings with later internalizing problems.

Epigenome-wide meta-analysis of prenatal maternal stressful life events and newborn DNA methylation.

Prenatal maternal stressful life events are associated with adverse neurodevelopmental outcomes in offspring. Biological mechanisms underlying these associations are largely unknown, but DNA methylation likely plays a role. This meta-analysis included twelve non-overlapping cohorts from ten independent longitudinal studies (N = 5,496) within the international Pregnancy and Childhood Epigenetics consortium to examine maternal stressful life events during pregnancy and DNA methylation in cord blood. Children whose mothers reported higher levels of cumulative maternal stressful life events during pregnancy exhibited differential methylation of cg26579032 in ALKBH3. Stressor-specific domains of conflict with family/friends, abuse (physical, sexual, and emotional), and death of a close friend/relative were also associated with differential methylation of CpGs in APTX, MyD88, and both UHRF1 and SDCCAG8, respectively; these genes are implicated in neurodegeneration, immune and cellular functions, regulation of global methylation levels, metabolism, and schizophrenia risk. Thus, differences in DNA methylation at these loci may provide novel insights into potential mechanisms of neurodevelopment in offspring.

Longitudinal associations of DNA methylation and sleep in children: a meta-analysis.

BACKGROUND: Sleep is important for healthy functioning in children. Numerous genetic and environmental factors, from conception onwards, may influence this phenotype. Epigenetic mechanisms such as DNA methylation have been proposed to underlie variation in sleep or may be an early-life marker of sleep disturbances. We examined if DNA methylation at birth or in school age is associated with parent-reported and actigraphy-estimated sleep outcomes in children. METHODS: We meta-analysed epigenome-wide association study results. DNA methylation was measured from cord blood at birth in 11 cohorts and from peripheral blood in children (4-13 years) in 8 cohorts. Outcomes included parent-reported sleep duration, sleep initiation and fragmentation problems, and actigraphy-estimated sleep duration, sleep onset latency and wake-after-sleep-onset duration. RESULTS: We found no associations between DNA methylation at birth and parent-reported sleep duration (n = 3658), initiation problems (n = 2504), or fragmentation (n = 1681) (p values above cut-off 4.0 × 10-8). Lower methylation at cg24815001 and cg02753354 at birth was associated with longer actigraphy-estimated sleep duration (p = 3.31 × 10-8, n = 577) and sleep onset latency (p = 8.8 × 10-9, n = 580), respectively. DNA methylation in childhood was not cross-sectionally associated with any sleep outcomes (n = 716-2539). CONCLUSION: DNA methylation, at birth or in childhood, was not associated with parent-reported sleep. Associations observed with objectively measured sleep outcomes could be studied further if additional data sets become available.

Connecting the dots: social networks in the classroom and white matter connections in the brain.

BACKGROUND: Peer connections in school classrooms play an important role in social-emotional development and mental health. However, research on the association between children's peer relationships and white matter connections in the brain is scarce. We studied associations between peer relationships in the classroom and white matter structural connectivity in a pediatric population-based sample. METHODS: Bullying and victimization, as well as rejection and acceptance, were assessed in classrooms in 634 children at age 7. White matter microstructure (fractional anisotropy (FA), mean diffusivity (MD)) was measured with diffusion tensor imaging at age 10. We examined global metrics of white matter microstructure and used Tract-Based Spatial Statistics (TBSS) for voxel-wise associations. RESULTS: Peer victimization was associated with higher global FA and lower global MD and peer rejection was associated with lower global MD; however, these associations did not remain after multiple testing correction. Voxel-wise TBSS results for peer victimization and rejection were in line with global metrics both in terms of direction and spatial extent of the associations, with associated voxels (pFWE <.05) observed throughout the brain (including corpus callosum, corona radiata, sagittal stratum and superior longitudinal fasciculi). CONCLUSIONS: Although based only on cross-sectional data, the findings could indicate accelerated white matter microstructure maturation in certain brain areas of children who are victimized or rejected more often. However, repeated measurements are essential to unravel this complex interplay of peer connections, maturation and brain development over time.

Maternal Mediterranean diet in pregnancy and newborn DNA methylation: a meta-analysis in the PACE Consortium.

Higher adherence to the Mediterranean diet during pregnancy is related to a lower risk of preterm birth and to better offspring cardiometabolic health. DNA methylation may be an underlying biological mechanism. We evaluated whether maternal adherence to the Mediterranean diet was associated with offspring cord blood DNA methylation.We meta-analysed epigenome-wide association studies (EWAS) of maternal adherence to the Mediterranean diet during pregnancy and offspring cord blood DNA methylation in 2802 mother-child pairs from five cohorts. We calculated the relative Mediterranean diet (rMED) score with range 0-18 and an adjusted rMED excluding alcohol (rMEDp, range 0-16). DNA methylation was measured using Illumina 450K arrays. We used robust linear regression modelling adjusted for child sex, maternal education, age, smoking, body mass index, energy intake, batch, and cell types. We performed several functional analyses and examined the persistence of differential DNA methylation into childhood (4.5-7.8 y).rMEDp was associated with cord blood DNA methylation at cg23757341 (0.064% increase in DNA methylation per 1-point increase in the rMEDp score, SE = 0.011, P = 2.41 × 10-8). This cytosine-phosphate-guanine (CpG) site maps to WNT5B, associated with adipogenesis and glycaemic phenotypes. We did not identify associations with childhood gene expression, nor did we find enriched biological pathways. The association did not persist into childhood.In this meta-analysis, maternal adherence to the Mediterranean diet (excluding alcohol) during pregnancy was associated with cord blood DNA methylation level at cg23757341. Potential mediation of DNA methylation in associations with offspring health requires further study.

Mapping gene by early life stress interactions on child subcortical brain structures: A genome-wide prospective study.

Background: Although it is well-established that both genetics and the environment influence brain development, they are typically examined separately. Here, we aimed to prospectively investigate the interactive effects of genetic variants-from a genome-wide approach-and early life stress (ELS) on child subcortical brain structures, and their association with subsequent mental health problems. Method: Primary analyses were conducted using data from the Generation R Study (N = 2257), including genotype and cumulative prenatal and postnatal ELS scores (encompassing life events, contextual risk, parental risk, interpersonal risk, direct victimisation). Neuroimaging data were collected at age 10 years, including intracranial and subcortical brain volumes (accumbens, amygdala, caudate, hippocampus, pallidum, putamen, thalamus). Genome-wide association and genome-wide-by-environment interaction analyses (GWEIS, run separately for prenatal/postnatal ELS) were conducted for eight brain outcomes (i.e., 24 genome-wide analyses) in the Generation R Study (discovery). Polygenic scores (PGS) using the resulting weights were calculated in an independent (target) cohort (adolescent brain cognitive development Study; N = 10,751), to validate associations with corresponding subcortical volumes and examine links to later mother-reported internalising and externalising problems. Results: One GWEIS-prenatal stress locus was associated with caudate volume (rs139505895, mapping onto PRSS12 and NDST3) and two GWEIS-postnatal stress loci with the accumbens (rs2397823 and rs3130008, mapping onto CUTA, SYNGAP1, and TABP). Functional annotation revealed that these genes play a role in neuronal plasticity and synaptic function, and have been implicated in neuro-developmental phenotypes, for example, intellectual disability, autism, and schizophrenia. None of these associations survived a more stringent correction for multiple testing across all analysis sets. In the validation sample, all PGSgenotype were associated with their respective brain volumes, but no PGSGxE associated with any subcortical volume. None of the PGS associated with internalising or externalising problems. Conclusions: This study lends novel suggestive insights into gene-environment interplay on the developing brain as well as pointing to promising candidate loci for future replication and mechanistic studies.

Epigenome-wide associations between observed maternal sensitivity and offspring DNA methylation: a population-based prospective study in children.

BACKGROUND: Experimental work in animals has shown that DNA methylation (DNAm), an epigenetic mechanism regulating gene expression, is influenced by typical variation in maternal care. While emerging research in humans supports a similar association, studies to date have been limited to candidate gene and cross-sectional approaches, with a focus on extreme deviations in the caregiving environment. METHODS: Here, we explored the prospective association between typical variation in maternal sensitivity and offspring epigenome-wide DNAm, in a population-based cohort of children (N = 235). Maternal sensitivity was observed when children were 3- and 4-years-old. DNAm, quantified with the Infinium 450 K array, was extracted at age 6 (whole blood). The influence of methylation quantitative trait loci (mQTLs), DNAm at birth (cord blood), and confounders (socioeconomic status, maternal psychopathology) was considered in follow-up analyses. RESULTS: Genome-wide significant associations between maternal sensitivity and offspring DNAm were observed at 13 regions (p < 1.06 × 10-07), but not at single sites. Follow-up analyses indicated that associations at these regions were in part related to genetic factors, confounders, and baseline DNAm levels at birth, as evidenced by the presence of mQTLs at five regions and estimate attenuations. Robust associations with maternal sensitivity were found at four regions, annotated to ZBTB22, TAPBP, ZBTB12, and DOCK4. CONCLUSIONS: These findings provide novel leads into the relationship between typical variation in maternal caregiving and offspring DNAm in humans, highlighting robust regions of associations, previously implicated in psychological and developmental problems, immune functioning, and stress responses.

Developmental Changes in Dynamic Functional Connectivity From Childhood Into Adolescence.

The longitudinal study of typical neurodevelopment is key for understanding deviations due to specific factors, such as psychopathology. However, research utilizing repeated measurements remains scarce. Resting-state functional magnetic resonance imaging (MRI) studies have traditionally examined connectivity as 'static' during the measurement period. In contrast, dynamic approaches offer a more comprehensive representation of functional connectivity by allowing for different connectivity configurations (time varying connectivity) throughout the scanning session. Our objective was to characterize the longitudinal developmental changes in dynamic functional connectivity in a population-based pediatric sample. Resting-state MRI data were acquired at the ages of 10 (range 8-to-12, n = 3,327) and 14 (range 13-to-15, n = 2,404) years old using a single, study-dedicated 3 Tesla scanner. A fully-automated spatially constrained group-independent component analysis (ICA) was applied to decompose multi-subject resting-state data into functionally homogeneous regions. Dynamic functional network connectivity (FNC) between all ICA time courses were computed using a tapered sliding window approach. We used a k-means algorithm to cluster the resulting dynamic FNC windows from each scan session into five dynamic states. We examined age and sex associations using linear mixed-effects models. First, independent from the dynamic states, we found a general increase in the temporal variability of the connections between intrinsic connectivity networks with increasing age. Second, when examining the clusters of dynamic FNC windows, we observed that the time spent in less modularized states, with low intra- and inter-network connectivity, decreased with age. Third, the number of transitions between states also decreased with age. Finally, compared to boys, girls showed a more mature pattern of dynamic brain connectivity, indicated by more time spent in a highly modularized state, less time spent in specific states that are frequently observed at a younger age, and a lower number of transitions between states. This longitudinal population-based study demonstrates age-related maturation in dynamic intrinsic neural activity from childhood into adolescence and offers a meaningful baseline for comparison with deviations from typical development. Given that several behavioral and cognitive processes also show marked changes through childhood and adolescence, dynamic functional connectivity should also be explored as a potential neurobiological determinant of such changes.

Neonatal DNA methylation and childhood low prosocial behavior: An epigenome-wide association meta-analysis.

Low prosocial behavior in childhood has been consistently linked to later psychopathology, with evidence supporting the influence of both genetic and environmental factors on its development. Although neonatal DNA methylation (DNAm) has been found to prospectively associate with a range of psychological traits in childhood, its potential role in prosocial development has yet to be investigated. This study investigated prospective associations between cord blood DNAm at birth and low prosocial behavior within and across four longitudinal birth cohorts from the Pregnancy And Childhood Epigenetics (PACE) Consortium. We examined (a) developmental trajectories of "chronic-low" versus "typical" prosocial behavior across childhood in a case-control design (N = 2,095), and (b) continuous "low prosocial" scores at comparable cross-cohort time-points (N = 2,121). Meta-analyses were performed to examine differentially methylated positions and regions. At the cohort-specific level, three CpGs were found to associate with chronic low prosocial behavior; however, none of these associations was replicated in another cohort. Meta-analysis revealed no epigenome-wide significant CpGs or regions. Overall, we found no evidence for associations between DNAm patterns at birth and low prosocial behavior across childhood. Findings highlight the importance of employing multi-cohort approaches to replicate epigenetic associations and reduce the risk of false positive discoveries.

Genome-wide DNA methylation patterns associated with general psychopathology in children.

Psychiatric symptoms are interrelated and found to be largely captured by a general psychopathology factor (GPF). Although epigenetic mechanisms, such as DNA methylation (DNAm), have been linked to individual psychiatric outcomes, associations with GPF remain unclear. Using data from 440 children aged 10 years participating in the Generation R Study, we examined the associations of DNAm with both general and specific (internalizing, externalizing) factors of psychopathology. Genome-wide DNAm levels, measured in peripheral blood using the Illumina 450K array, were clustered into wider co-methylation networks ('modules') using a weighted gene co-expression network analysis. One co-methylated module associated with GPF after multiple testing correction, while none associated with the specific factors. This module comprised of 218 CpG probes, of which 198 mapped onto different genes. The CpG most strongly driving the association with GPF was annotated to FZD1, a gene that has been implicated in schizophrenia and wider neurological processes. Associations between the probes contained in the co-methylated module and GPF were supported in an independent sample of children from the Avon Longitudinal Study of Parents and Children (ALSPAC), as evidenced by significant correlations in effect sizes. These findings might contribute to improving our understanding of dynamic molecular processes underlying complex psychiatric phenotypes.

DNA methylation signatures of aggression and closely related constructs: A meta-analysis of epigenome-wide studies across the lifespan.

DNA methylation profiles of aggressive behavior may capture lifetime cumulative effects of genetic, stochastic, and environmental influences associated with aggression. Here, we report the first large meta-analysis of epigenome-wide association studies (EWAS) of aggressive behavior (N = 15,324 participants). In peripheral blood samples of 14,434 participants from 18 cohorts with mean ages ranging from 7 to 68 years, 13 methylation sites were significantly associated with aggression (alpha = 1.2 × 10-7; Bonferroni correction). In cord blood samples of 2425 children from five cohorts with aggression assessed at mean ages ranging from 4 to 7 years, 83% of these sites showed the same direction of association with childhood aggression (r = 0.74, p = 0.006) but no epigenome-wide significant sites were found. Top-sites (48 at a false discovery rate of 5% in the peripheral blood meta-analysis or in a combined meta-analysis of peripheral blood and cord blood) have been associated with chemical exposures, smoking, cognition, metabolic traits, and genetic variation (mQTLs). Three genes whose expression levels were associated with top-sites were previously linked to schizophrenia and general risk tolerance. At six CpGs, DNA methylation variation in blood mirrors variation in the brain. On average 44% (range = 3-82%) of the aggression-methylation association was explained by current and former smoking and BMI. These findings point at loci that are sensitive to chemical exposures with potential implications for neuronal functions. We hope these results to be a starting point for studies leading to applications as peripheral biomarkers and to reveal causal relationships with aggression and related traits.

Maternal anxiety during pregnancy and newborn epigenome-wide DNA methylation.

Maternal anxiety during pregnancy is associated with adverse foetal, neonatal, and child outcomes, but biological mechanisms remain unclear. Altered foetal DNA methylation (DNAm) has been proposed as a potential underlying mechanism. In the current study, we performed a meta-analysis to examine the associations between maternal anxiety, measured prospectively during pregnancy, and genome-wide DNAm from umbilical cord blood. Sixteen non-overlapping cohorts from 12 independent longitudinal studies of the Pregnancy And Childhood Epigenetics Consortium participated, resulting in a combined dataset of 7243 mother-child dyads. We examined prenatal anxiety in relation to genome-wide DNAm and differentially methylated regions. We observed no association between the general symptoms of anxiety during pregnancy or pregnancy-related anxiety, and DNAm at any of the CpG sites, after multiple-testing correction. Furthermore, we identify no differentially methylated regions associated with maternal anxiety. At the cohort-level, of the 21 associations observed in individual cohorts, none replicated consistently in the other cohorts. In conclusion, contrary to some previous studies proposing cord blood DNAm as a promising potential mechanism explaining the link between maternal anxiety during pregnancy and adverse outcomes in offspring, we found no consistent evidence for any robust associations between maternal anxiety and DNAm in cord blood. Larger studies and analysis of DNAm in other tissues may be needed to establish subtle or subgroup-specific associations between maternal anxiety and the foetal epigenome.

Epigenome-wide change and variation in DNA methylation in childhood: trajectories from birth to late adolescence.

DNA methylation (DNAm) is known to play a pivotal role in childhood health and development, but a comprehensive characterization of genome-wide DNAm trajectories across this age period is currently lacking. We have therefore performed a series of epigenome-wide association studies in 5019 blood samples collected at multiple time-points from birth to late adolescence from 2348 participants of two large independent cohorts. DNAm profiles of autosomal CpG sites (CpGs) were generated using the Illumina Infinium HumanMethylation450 BeadChip. Change over time was widespread, observed at over one-half (53%) of CpGs. In most cases, DNAm was decreasing (36% of CpGs). Inter-individual variation in linear trajectories was similarly widespread (27% of CpGs). Evidence for non-linear change and inter-individual variation in non-linear trajectories was somewhat less common (11 and 8% of CpGs, respectively). Very little inter-individual variation in change was explained by sex differences (0.4% of CpGs) even though sex-specific DNAm was observed at 5% of CpGs. DNAm trajectories were distributed non-randomly across the genome. For example, CpGs with decreasing DNAm were enriched in gene bodies and enhancers and were annotated to genes enriched in immune-developmental functions. In contrast, CpGs with increasing DNAm were enriched in promoter regions and annotated to genes enriched in neurodevelopmental functions. These findings depict a methylome undergoing widespread and often non-linear change throughout childhood. They support a developmental role for DNA methylation that extends beyond birth into late adolescence and has implications for understanding life-long health and disease. DNAm trajectories can be visualized at http://epidelta.mrcieu.ac.uk.

Genome-wide DNA methylation patterns associated with sleep and mental health in children: a population-based study.

BACKGROUND: DNA methylation (DNAm) has been implicated in the biology of sleep. Yet, how DNAm patterns across the genome relate to different sleep outcomes, and whether these associations overlap with mental health is currently unknown. Here, we investigated associations of DNAm with sleep and mental health in a pediatric population. METHODS: This cross-sectional study included 465 10-year-old children (51.3% female) from the Generation R Study. Genome-wide DNAm levels were measured using the Illumina 450K array (peripheral blood). Sleep problems were assessed from self-report and mental health outcomes from maternal questionnaires. Wrist actigraphy was used in 188 11-year-old children to calculate sleep duration and midpoint sleep. Weighted gene co-expression network analysis was used to identify highly comethylated DNAm 'modules', which were tested for associations with sleep and mental health outcomes. RESULTS: We identified 64 DNAm modules, one of which associated with sleep duration after covariate and multiple testing adjustment. This module included CpG sites spanning 9 genes on chromosome 17, including MAPT - a key regulator of Tau proteins in the brain involved in neuronal function - as well as genes previously implicated in sleep duration. Follow-up analyses suggested that DNAm variation in this region is under considerable genetic control and shows strong blood-brain concordance. DNAm modules associated with sleep did not overlap with those associated with mental health. CONCLUSIONS: We identified one DNAm region associated with sleep duration, including genes previously reported by recent GWAS studies. Further research is warranted to examine the functional role of this region and its longitudinal association with sleep.