Proposal summaries
B1497 - Facial dysmorphism and maternal alcohol consumption in the ALSPAC cohort - 31/01/2013
Background
Development of the face is dependent on complex interactions of genetic and environmental factors, and is also affected by maternally derived exposures. Alcohol consumption in pregnancy is an important health issue and the DOH currently advise that pregnant women or women trying to conceive should avoid drinking alcohol1.Prenatal alcohol exposure causes a continuum of effectsand is thought to be a leading cause of learning disability in the western world with between 0.2 and 2.0 cases per 1,000 live births. Some of the highest incidence levels have been identified in Eastern Europe, in indigeneous North American communities, and in the South African Cape-Coloured population.The most severe phenotype, fetal alcohol syndrome (FAS), affects face shape, growth and neurobehavior. Fetal alcohol spectrum disorders (FASD) include FAS and other pathologies arising from prenatal alcohol exposure.
The characteristic facial features believed to be due to alcohol ingestion between 10 and 20 wks of gestation include smooth philtrum, thin upper lip vermilion, long upper lip, shortened palpebral aperture width, flattened mid-face and small head circumference. The timing of maternal alcohol consumption is considered to affect the severity of facial dysmorphology and through PH's involvement in the CIFASD consortium (www.CIFASD.com), we recently reported on the link between neurofacial effects and timing of ethanol exposure in a mouse model2. In another CIFASD study, of South African children3, we used face shape to induce classification schemes and tested agreement with clinical FASD categorization: FAS, partial FAS, and HE (heavily alcohol exposed without clinically detectable facial features). The more heterogeneous phenotype of HE forced us to introduce a novel clustering technique, signature graph analysis4, which normalizes face shape and links individuals with similar facial dysmorphism. Signature graph analysis identified half of the HE group as having facial dysmorphism more FAS-like than control-like. These HE individuals performed less well on psychometric tests than HE individuals who facially were more control-like. We also demonstrated that heat map comparisons of, and animated morphs between, individual faces and matched control means revealed facial dysmorphism otherwise overlooked. Thus, visualizations and signature analysis can help pediatricians detect facial dysmorphism across the fetal alcohol spectrum, especially in non-syndromal alcohol exposed cases.
The ALSPAC database provides an opportunity for a much larger-scale investigation of associations between volume and timing of maternal alcohol consumption during pregnancy and facial dysmorphism. In addition, we wish to explore further associations between facial characteristics, cognitive impairment and psychosis
Subjects and data
There are 4747 3D facial scans captured for the 15 year old cohort and these images provide detailed facial morphology at tens of thousands of 3D surface points. The 3D scans have been landmarked and derived facial features recorded. ALSPAC has also collected information on maternal alcohol consumption through questionnaires at 12, 18 and 32 weeks of gestation as well as data regarding the drinking consumption of the parents. In addition, quantity of maternal alcohol intake data is available related to the timing of fetal development: before pregnancy, first trimester of pregnancy, and at around the time the baby was first felt to move. The dataset identified above will need to be extracted and sent to both Peter Hammond and Stephen Richmond.
B1493 - Utilising population-based collections from the UK to identify genetic risk factors for idiopathic scoliosis - 31/01/2013
Aims
We aim to carry out the first population-based genome wide association studies for the presence of scoliosis using white European populations, and to investigate whether similar genetic associations are seen in different ethnic groups. In addition, by utilising resources already available from ALSPAC we will carry out preliminary functional investigations on any identified genetic loci.
To utilise the genetic data already collected to investigate the genetic contributions to scoliosis we will
1. Investigate if the SNP rs11190870 is associated with the presence of scoliosis
2. Perform GWAS in ALSPAC
3. Meta-analysis of GWAS from ALSPAC and Twins-UK
4. Evaluate if identified novel SNPs are likely to have biological relevance bioinformatically and by utilising gene expression, DNA methylation and metabolomic data
5. Continue to develop a model to predict the progression of scoliosis
Methods 1: GWAS
Only those common genotypic variants with a minor allele frequency greater than 5% will be studied. Only SNPs which passed an exact test of Hardy-Weinberg equilibrium are considered for analysis. Association analysis will be performed using logistic regression models based on log additive models, adjusting for age, gender and other appropriate variables. We will assess genome-wide data for associations with scoliosis in ALSPAC alone, setting genome-wide significance at P<=5x10-8.
Methods 2: Power
The power of our GWAS depends on the minor allele frequency of SNPs, the likely size of effect and our cut-off for defining scoliosis. Given the two previous genome-wide studies have found an effect size of 1.56 and 1.85, and the rs11190870 has an allele frequency of 0.437 in ALSPAC, we have reasonable power to detect de novo effects of a comparable magnitude to those previously published.
Methods 3: Meta-analysis
We will perform a meta-analysis of the ALSPAC and Twins-UK GWAS. Prior to meta-analysis, poorly imputed SNPs and those with an allele frequencyless than 5% will be excluded. Inverse variance fixed-effects meta-analyses will be undertaken using METAL. Genomic control corrections will be applied before reporting SNPs which reach genome-wide significance (Pless than 1x10-5) for further investigation.
Methods 4: Replication
We will select signal loci (+/-500mb) with the greatest evidence for association with the risk of scoliosis for replication within data from Japanese and Hong-Kong case-control studies that have GWAS data already available. Population linkage disequilibrium (LD) will be taken into account and where possible used to aid fine mapping. The Chinese Hong Kong disease cohort will then be genotyped for our top 10 SNP hits. Associations between SNPs and scoliosis will then be analysed as described above. We will then repeat the meta-analysis based on all five cohorts.
Methods 5: Bioinformatics
We will use all existing knowledge to analyze our identified SNPs in order to establish its status as a potential functional variant for scoliosis. A bioinformatic approach will be taken to help identify which gene the signal is in, and its likely functional networks. The potential impact of coding variation will be assesed with a series of predictive approaches including SIFT and PolyPhen and we will also followup non-coding functional elements recently annotated by the ENCODE consortium. This will generate prioritised lists of variants for further functional examination in future research projects.
Methods 6: Integrated use of available expression, methylation and metabolomic data
We will examine the impact of our identified SNPs on patterns of protein expression in cell lines, in DNA methylation and through examination of detailed banks of metabolomic data. Along with a sub-set of ALSPAC and Twins-UK samples with transformed multi-tissue specific expression data, the BBSRC-funded ARIES study provides an opportunity to examine the likely effect of our identified SNPs on methylation, as a potential mechanism of gene-environment interaction. This will allow us to extend our primary goal and to begin to unpick the contributing nature of genetic loci to scoliosis risk in a unique study environment.
Exposure variables
Common genotypic variants genotyped using the Illumina HumanHap550 platform.
Outcome variable
Scoliosis (yes/no as a binary variable) identified by the DSM at aged 9 and 15, will be defined as those with a curve >=10degrees, as this has substantial repeatability (Kappa of 0.74). Sensitivity analyses will be carried out using a lower cut-off of >=6degrees to define scoliosis, as the DSM underestimates curve size by approximately 40%, although this lower cut-off has only moderate repeatability (Kappa of 0.56). In addition, we will explore the use of angle size as a continuous variable.
Confounding and other variables
Age and gender
Gene expression data
DNA methylation data
Metabolomic data
B1492 - Explaining risk for suicidal behaviour in adolescent offspring of depressed parents - 31/01/2013
AIMS
(1) Examine in detail the longitudinal association between maternal and paternal depression and the course of offspring suicidal behaviour.
(2) Assess the extent to which prior child psychopathology accounts for excess risk of suicidal behaviour in offspring of depressed parents relative to offspring of non-depressed parents.
(3) Assess the extent to which other hypothesised mediating mechanisms (such as family factors, peer problems and other child factors) account for excess risk of suicidal behaviour in offspring of depressed parents and to examine how these factors interact with child psychopathology.
(4) Where possible, use repeated measures to test the direction of effects for identified mediating mechanisms and adolescent suicidality.
B1503 - Mapping and assessing TP53-associated linkage disequilibrium blocks in cancer reproductive traits and viral infections - 31/01/2013
The general focus of the work to be carried out is to understand the importance of the TP53 gene and its product (p53 protein) to several aspects of human health and disease (comprehensively available in the ALSPAC cohort database), including cancer, reproductive traits (such as pregnancy loss and success of assited reproduction) and viral infections. The p53 protein is essentially a transcription factor that is well-understood to play crucial roles in cancer. This protein mediates the response to DNA damage signals by activating cellular pathways that result in senescence or apoptosis. Mutations in the coding region of TP53 have been found in approximately 50% of all human cancers (although this can vary considerably depending on the cancer type), thus being a hallmark of tumourigenesis. More recently, p53 has been evidenced to participate of other biological contexts, including skin colour, embryo implantation to the uterus, metabolic disorders, ageing and viral infections.
Germline genomic variations (mainly Single Nucleotide polymorphism - SNPs) in TP53 have thus been extensively studied as an attempt to identify genetic markers of cancer predisposition, which would represent a contribution to the clinical field. SNPs are, in general, low-penetrance genetic variations that depend on complex interactions with other genetic profiles and environmental factors to be associated with a certain phenotype or disease. Considering that SNPs that change protein primary amino acid sequence (non-synonymous SNPs - nsSNPs) tend to have more prominent effects on protein function than synonymous or non-coding SNPs (although the relevance of these two last have been evidenced in the context of mRNA expression, splicing or translation), the majority of the currently available association studies has given attention to TP53 nsSNPs. Although there is a considerable amount of nsSNPs reported for TP53, the current work focus is on four of them, which may be considered the best validated: P47S, R72P, V217M and G360A. The two first nsSNPs have been substantially evidenced to functionally impact p53 function (p53 transactivation activity and apoptosis efficiency, respectively) and have been associated with cancer risk, while the remaining two currently lack such association findings. Interestingly, both these nsSNPs have functional experiments suggesting they might be important in cancer (V217M is located in the DNA binding domain of p53 and G360A has been evidenced to impact the transactivation of some p53-interacting proteins).
The notion that the association between a SNP and a phenotype is complex and the experimental evidences for assuming V217M and G360 SNPs might have significant impacts on p53 function, the current project proposes the investigation of TP53 haplotypes (set of SNPs on a single chromosome that are statistically associated) regarding the four nsSNPs described, as well as linkage disequilibrium (LD) blocks (which are composed of genetic loci that do not - in a statistical sense - do not segregate randomly) with SNPs in other genes. For instance, the association of TP53 R72P SNP and response to UV exposure (and, consequently, associations with tanning and skin cancer) are dependent on SNPs in the MC1R gene (a typical epistatic effect).
Studies of this nature are particularly suitable for large population cohorts such as ALSPAC, since the tendency of reducing the number of individuals per genetic subgroup when analyzing haplotypes would not result in insufficient statistical power. Moreover, there are comprehensive data available regarding ALSPAC mothers health, which is invaluable for the present study. Since TP53 R72P SNP is well-established to be associated with skin color, ALSPAC data is also of interest because it has little ethnic admixture, reducing bias. In this topic, the large and high-quality characterization of the sample (including massive data at the genetic level) allows the construction and assessment of LD blocks with high accuracy, converging information of different pathways at the genetic level.
The aims of the project can be divided in: 1) Identify haplotype and LD blocks involving the nsSNPs of interest. 2) Investigate possible associations of these blocks with different disease traits available for ALSPAC cohort. 3) Investigate possible interactions of TP53 nsSNPs with SNPs in LD with them regarding the associations of the latter with traits currently not described as affected by TP53. 4) Understand how TP53-associated SNPs interact with each other in order to provide substantial insights related to the genetics of complex traits. ((5) With a possibel amendment, extend analyses to other variants collected in ALSPAC (from GWAS/sequence data) across the TP53 locus).
B1502 - Variations in wound healing loci and their association with injury and scar related outcomes in the ALSPAC cohort - 31/01/2013
Where adult tissue is damaged, a complex repair process is taking place involving regeneration and acute phase immunological response. Unlike embryonic tissues, adult repair always leads to the formation of a fibrotic scar where the wound has healed, which ultimately can disable proper tissue function [1]. In recent years, research was able to link several genes to the event of scar formation . Knockdown of Ostepontin (OPN) in mice for example resulted in reduced granulation tissue formation and scarring [2]. It also has been indicated that TGF-beta1 in conjunction with Connective tissue growth factor (CTGF) is promoting scar formation [3]. Most of this data comes from mouse model studies, in humans however, less is known.
Using ALSPAC data we want to perform both a candidate driven analysis and a non-hypothesis driven GWAS (the latter being determined by the sample sizes generated from available phenotypic data) comparing individuals involved in an accident developing a scar compared to individuals involved in an accident who did not develop a scar. Where possible, we will attempt to assess differing types of scar
tissue and healing related phenotype, however this will again be contingent upon available phenotypic data).
Analysis plan:
(i) assess the depth of data pertinent to scarring phenotypes and establish cases
control series according to differential scarring patterns.
(ii) perform bioinformatic work up of select genes (LIST) to establish likely
functional variants across the coding region and surrounding region.
(iii) unite both genetic and phenotypic data to undertake tests of association
between genetic variation and phenotypic characterisation.
B1501 - NUTRIMENTAL call under FP7-KBBE-2013-7 - 31/01/2013
The aim is to examine the role of life styles, social background and socioeconomic factors on the mood-nutrition reciprocal relationships using samples from countries with different diets and different levels of mental health problems. For example, Mediterranean, Scandinavian and UK or US samples have been investigated in isolation but no study has carried out a comparison of social and dietary influences on mental health in different countries. Other factors which can be examined are changes in diet-mental health relationships across different age groups, gender and socio-economic status.
The research methodology will be largely based on secondary analyses of existing databases. In some cases the diet-mental health relationships have already been examined and the novel part will be the inclusion of social influences into the analyses. However, there are also extensive new samples which can provide new information on these topics. For example, in the UK the Biobank initiative has led to the collection of relevant data from 350,000 individuals. This database is potentially available (for a small fee) to the present project.
B1500 - External validation of an algorithm to estimate overweight risk in childhood from predictors during infancy - 31/01/2013
Background:
In the United Kingdom in 2010, around three in ten boys and girls (aged 2 to 15) were classed as either overweight or obese. However, overweight and obesity rates are not currently available for the under 2s despite evidence that some infants are larger than desired. Estimates vary but it is thought that between 25% to 33% of infants gain weight rapidly during the first six months of life. Rapid weight gain during infancy is associated with obesity between 6- 8 years of age and later life. There is evidence that weight at 5 years of age is a good indicator of the future health of a child and that obesity during childhood increases the risk of adult obesity. This has a clearly measurable impact on physical and mental health, quality of life, and generates considerable direct and indirect costs. Thus, there is a compelling rationale for identifying those infants at greatest risk.
Previous work:
We have developed a risk score algorithm for overweight in childhood based on a predictor model in infants using the Millenium Cohort Study, a large British birth cohort. Stepwise logistic regression was used to determine a predictor model for childhood overweight at 3 years defined by the IOTF. A risk algorithm was developed by assigning integer values to beta-coefficients based on relative strength. Discrimination was analysed using the receiver operating characteristic (ROC) curve.The strongest predictors of overweight risk at 3 years were rapid weight gain during the first year of life, infant birth weight greater than 3.81 kg, maternal pre-pregnancy BMI from 25 kg/m^2 to 30 kg/m^2 and maternal pre-pregnancy BMI >= 30 kg/m^2. The total risk score ranged from a minimum of 0 to a maximum of 33 corresponding to a predicted risk of overweight from 5.1% to 68.8%. The c-statistic from the ROC was 0.70. While the risk algorithm has been proven to have good internal validity, we have yet to test/validate the algorithm on an external cohort.
Aim of current study:
Therefore, the aim of the current study is to determine the validity of the risk algorithm developed from the MCS using the ALSPAC data set.
Methods:
We have a developed risk equations from the MCS that we will apply to the infants in the ALSPAC cohort. We will compare the predictive risk of overweight at 3 years to the observed risk of overweight at 3 years. Discrimination will be assessed using ROC c-statistics (AUC). This will allow us to calibrate the existing risk equations if neccessary.
Key exposure variables:
Child's gender, household income, infant birth birth, infant weight gain the first year of life, maternal pre-pregnancy BMI, maternal smoking in pregnancy, infant breastfeeding status in the first year, formal childcare arrangements in the first year
Outcome variables: Child's weight at 3 years, Child's height at 3 years, Child exact age at follow-up
Final outcome:
Using those three variables, the primary clinical outcome for childhood overweight at 3 years was defined by the International Obesity Task Force sex and age-specific cut-offs corresponding to an adult BMI >= 25 kg/m^2 (Girls: 18.02 kg/m^2; Boys: 18.41 kg/m^2).
Other confounding variables for the investigation and calibration of the risk equations (We have conducted a systematic review which has informed variable selection):
Child's ethnicity, maternal marital status, maternal education, maternal employment, delivery type, maternal age, maternal alcohol consumption, maternal depression, diabetes, breastfeeding duration, formula feeding, age at introduction of solid food, infant temperment, feeding/eating, sleep, activity, SES
B1499 - Efficient estimation of causal treatment effects using genotypic data - 31/01/2013
Aims: We will investigate the optimal methods for combining multiple genetic variants to estimate the causal effects of risk factors on outcomes. This is a methodological investigation, however as an example, we will investigate the associations of height and weight and IQ, behavioural problems (hyperactivity, emotional problems, conduct problems and peer problems) and academic achievement across childhood and adolescence.
Mendelian randomization uses genetic variants as instrumental variables for modifiable risk factors.5 It has been widely applied to investigate the effects of risk factors for disease such as weight, blood pressure, cholesterol, and behavioural risk factors such as alcohol and tobacco consumption on health and socio-economic outcomes.1-4 To be valid instruments, variants must be associated with risk factors of interest, and have no direct effect on the outcome of interest. The stronger the association between the instrument and the exposure, the more statistical power and precision a Mendelian randomization analysis can achieve. One challenge in using genetic instrumental variables is that many variants only have modest effects on the risk factor of interest.6 The use of multiple variants as instruments, thereby explaining more of the variability in the risk factor,7 could therefore increase the precision of the results, improving the likelihood of being able to draw inferences from any given dataset.
Researchers have used various approaches to construct instruments, including allele scores, weighted allele scores, using each genetic variant as an independent instrument, and using generalized method of moment estimators to efficiently weight each of the variants. However, we currently do know the optimal approach for aggregating this information for Mendelian randomization analysis. Thus in this project we will investigate the optimal methods for combining multiple variants to maximise statistical power.
We will examine the effects of anthropometry on cognitive and behavioural outcomes as a motivating example. However our major objective is to develop methodologies rather than investigating any single hypotheses.
Hypotheses: What is the most efficient method for combining multiple variants into a Mendelian randomization analysis?
Exposures: Height, weight and adiposity at age 8-13, genome-wide data.
Outcomes: Academic achievement, IQ score aged 8, measures at age 13 of hyperactivity, emotional problems, conduct problems, and peer problems (from SDQ).
Potential confounding variables or descriptive data: Age, birth weight, number of younger and older siblings, income, parents' education, social class and employment status. Index of multiple deprivation of families' neighbourhood. Mothers' characteristics during pregnancy: age, locus of control, EPDS, CCEI, and teaching score.
B1498 - MSc project Validation of essential fatty acid intakes using FFQ at age 7 with plasma fatty acid measures - 31/01/2013
MSc Project for Nutritions student in School for Policy studies
We have been unable to calculate the intake of essential fatty acids from the FFQ to date due to lack of funding - this is an oportunity to obtain this useful data and partially validate it against blood fatty acid levels at age 7 years.
The student would be provided with a copy of the macro used to calculate the intake of total polyunsaturated fatty acids (PUFA) from the food frequency questionnaire completed by parents for their child age x months (7y FFQ). We would also provide a list of the foods and portion sizes used to calculate total PUFA. The student would search for data on the content of selected fatty acids, particularly Arachadonic acid and Alpha-linoleic acid (the precursors of the omega-6 and omega-3 series respectively) of the foods used. They would record the sources of all data on a spread sheet and complete the macro to enable calculation of each of these new fatty acids.
The calculation of the intake of the fatty acids from ALSPAC participants data would be carried out by Pauline Emmett so that the student would not need access to the ALSPAC dataset. Unlinked summary results would be used by the student to assess possible outliers and for their MSc write-up. The fatty acid intakes thus claculated would be correlated (by Pauline Emmett) with blood measures of these fatty acids to assess the effectiveness of the FFQ to assess these nutrients. Summary unlinked results would be returned to the student. The calculated fatty acid intake for ALSPAC 7 year olds would become available to other users at the end of this project.
B1490 - Educational attainment and BMI bidrectional assessment of causal relationships using Mendelian randomisation - 17/01/2013
he aim of this miniproject is to look for relationships between educational attainment and BMI and assess causality between them using Mendelian randomisation. This will involve:
- Generating educational attainment (EA) scores derived from genotypic associations with EA from an independent study (carried out as part of the Social Science Genetics Association Consortium) and genotypic dosage information within ALSPAC data.
- Assessing the performance of a series of genotype based scores of EA in terms of their predictive ability for KS3 SATS results (and IQ) and their performance as potential instruments for EA within Mendelian randomisation frameworks.
- Undertaking both observational assessment and Mendelian randomisation analysis of relationships between EA and child BMI.
- Generating similar genetic scores for BMI (initially derived from the 32 most predictive loci taken from Speliotes et al, 2010) and undertake the same analyses as above, but in a bidirectional manner.
In order to perform these analyses, we will require access to ALSPAC genetic (I already have access via the alspac-shared directory on BlueCrystal) data and a series of key phenotypes outlined in the table earlier in this form.
Reference:
Speliotes et al (2010) Association analyses of 249,796 individuals reveal 18 new loci associated with body mass index. Nature Genetics. 42 (11). pp. 937-948.
B1489 - Associations between primary care measures and psychosis-like symptoms in the ALSPAC birth cohort - 17/01/2013
Aim:
The aim of this research is to identify if associations are present between PLIKS, recorded within ALSPAC, and primary health care measures, taken from GPRD.
B1488 - Maternal alcohol use during pregnancy and offspring growth trajectories - 17/01/2013
Maternal alcohol drinking during pregnancy and offspring trajectories of height, weight and head circumference
In Denmark, United Kingdom, Australia and Ireland it has been estimated that between 37% and 81% of fetuses are exposed to alcohol in pregnancy (1-5). While the hazards of heavy alcohol consumption in pregnancy on birth outcomes including preterm birth and birth weight are well recognized, the effects of moderate-low levels of alcohol drinking in pregnancy are less clear both at birth and in particular longitudinally. The association between alcohol use during pregnancy and postnatal growth throughout childhood is not well established. Studies which focus on postnatal growth are a priority (6, 7) if consensus on safe alcohol recommendations for pregnant women is to be reached.
Aim : To investigate whether different patterns of alcohol use during pregnancy adversely affect height, weight and head circumference between birth and 10 years
Objective #1 Recent work conducted using the ALSPAC data investigated maternal smoking during pregnancy and offspring trajectories of height and adiposity up to age 10 (8). We propose to replicate the approach taken in this analysis in an investigation of the associations between different levels of alcohol consumption during pregnancy and height, weight and head circumference trajectories between birth and age 10 years.
Briefly, we will estimate the association between maternal drinking during pregnancy (any versus none and by dose) and trajectories of height, weight and head circumference that have previously been developed using multilevel models by Laura Howe and Andrew Smith. In keeping with Lewis and colleagues' recent categorisation of alcohol use in pregnancy (9) moderate alcohol consumption will be defined as equal to 1-6 drinks on average per week during pregnancy, heavy consumption will be defined as consumption of greater than 6 drinks per day and binge drinkers will be classified as women who report consuming more than 4 units of alcohol at 18 and 32 weeks of pregnancy. This classification will be examined with reference to the growth trajectories of women who abstained from alcohol during pregnancy.
Covariates adjusted for will include maternal age, ethnicity, maternal education, household occupational class, parity, maternal height, paternal height, maternal BMI, smoking during pregnancy, and other key variables. We will also consider drug use as a potential confounder, but this has a very low prevalence, so we will investigate whether or not it is possible/appropriate to include in analyses.
Objective #2 Residual confounding is often a significant limitation in studies of fetal alcohol exposure and infant outcomes. The relationship between alcohol use during pregnancy and growth outcomes at birth could be confounded by shared familial characteristics as has been suggested for the relationship between smoking during pregnancy and growth (8). For this part of the analysis, we will compare the associations of maternal alcohol consumption during pregnancy to the woman's partners in order to explore the presence of unmeasured confounders. This will allow us to estimate the extent to which any associations detected are likely to be the result of unmeasured genetic, socioeconomic or behavioural confounders.
Partner alcohol consumption was collected in a questionnaire sent to them at 18 weeks gestation. Partner alcohol consumption will be classified as moderate, heavy and/or binge in line with UK recommendations of no more than 21 units per week and no more than 4 units in one occasion. We will run the analysis specified in objective #1 both with mutual adjustment for partner alcohol consumption and for partner consumption alone adjusting for covariates already specified.
As a further tool for the assessment of potential unmeasured confounding, we will compare the growth trajectories of children whose mothers drink alcohol during pregnancy with those who do not drink during pregnancy but resume alcohol consumption shortly after delivery.
References
1. Colvin L, Payne J, Parsons D, Kurinczuk JJ, Bower C. Alcohol consumption during pregnancy in nonindigenous west Australian women. Alcoholism: Clinical and Experimental Research. 2007;31(2):276-84.
2. Kelly Y, Sacker A, Gray R, Kelly J, Wolke D, Quigley MA. Light drinking in pregnancy, a risk for behavioural problems and cognitive deficits at 3 years of age? International Journal of Epidemiology. 2009;38(1):129-40.
3. Bakker R, Pluimgraaff LE, Steegers EAP, Raat H, Tiemeier H, Hofman A, et al. Associations of light and moderate maternal alcohol consumption with fetal growth characteristics in different periods of pregnancy: The Generation R Study. International Journal of Epidemiology. 2010;39(3):777-89.
4. Andersen AMN, Andersen PK, Olsen J, Gronbaek M, Strandberg-Larsen K. Moderate alcohol intake during pregnancy and risk of fetal death. International Journal of Epidemiology. 2012.
5. Mullally A, Cleary BJ, Barry J, Fahey TP, Murphy DJ. Prevalence, predictors and perinatal outcomes of peri-conceptional alcohol exposure - retrospective cohort study in an urban obstetric population in Ireland. BMC Pregnancy and Childbirth. 2011;11.
6. Patra J, Bakker R, Irving H, Jaddoe V, Malini S, Rehm J. Dose-response relationship between alcohol consumption before and during pregnancy and the risks of low birthweight, preterm birth and small for gestational age (SGA)-a systematic review and meta?analyses. BJOG: An International Journal of Obstetrics & Gynaecology.
7. Henderson J, Gray R, Brocklehurst P. Systematic review of effects of low-moderate prenatal alcohol exposure on pregnancy outcome. BJOG: An International Journal of Obstetrics & Gynaecology. 2007;114(3):243-52.
8. Howe LD, Matijasevich A, Tilling K, Brion MJ, Leary SD, Smith GD, et al. Maternal smoking during pregnancy and offspring trajectories of height and adiposity: comparing maternal and paternal associations. International journal of epidemiology. 2012;41(3):722-32.
9. Lewis SJ, Zuccolo L, Smith GD, Macleod J, Rodriguez S, Draper ES, et al. Fetal Alcohol Exposure and IQ at Age 8: Evidence from a Population-Based Birth-Cohort Study. PloS one. 2012;7(11):e49407.
B1487 - BRAINEAT - 17/01/2013
There is a bidirectional relationship between the brain and nutrition. The brain influences our food
choices and, therefore, the amount of macro- and micro-nutrients entering our bodies. These nutrients
affect body composition and a variety of metabolic processes that, in turn, influence brain function and
structure. Understanding the various elements of this brain-body loop is one of the overarching goals of
the proposed research, both in terms of gaining new knowledge about the underlying processes and
applying it to enhance population health.
We will take advantage of several community-based cohorts including ALSPAC to investigate bidirectional
brain-body relationships in the context of food choices; magnetic resonance images (MRIs) of
the brain and, in most cases, of abdominal fat have been collected in these cohorts together with a wealth
of other relevant information (e.g., food recall interviews) and blood samples suitable for genetic and
metabolomics analyses.
1. Cohorts
Given the high cost of acquiring large samples with imaging-based, systems-level phenotypes, we have
brought together several cohorts in which MRI of the brain and body have been already accomplished or
funded (~8,000 participants). Three of these samples are birth cohorts (ALSPAC, Generation R, and
NFBC86), thus allowing us to use all of the longitudinal data collected so far. In addition, we will be able
to incorporate the (Canadian) Saguenay Family Study, which is unique with regards to its twogenerational
design (adolescents and their middle-aged parents) and the detailed cardiovascular and
metabolic phenotyping available in all participants (15 hours of assessments).
Using these datasets, we will ask: (1) how do inter-individual differences in the structural (and functional)
properties of the relevant neural circuits relate to food choices, such as fat or carbohydrate intake; and (2)
how do inter-individual differences in body composition (visceral fat from MRI and/or body composition
from bioimpedance) influence metabolic profiles (as determined by metabolomics) and, in turn, brain
structure and function.
2. Magnetic resonance imaging
All cohorts acquired T1-weighted images of the brain, which are well suited for the quantification of a
number of grey-matter properties (see below). Furthermore, all cohorts used MR sequences (Diffusion
Tensor Images [DTI] and/or Magnetization Transfer Images [MTI]) suitable for quantifying various
properties of white matter. Three of the cohorts (Generation-R, NFBC86, and the parent arm of the SFS)
have acquired resting-state functional MRI (fMRIrs), and one of the cohorts (ALSPAC) has acquired a
paradigm-based fMRI (viewing faces; Grosbras and Paus 2006).
Using this rich dataset, we plan to use image-processing pipelines developed and implemented at two of
the sites (Erasmus University and University of Toronto) to process these images in order to generate
quantitative phenotypes suitable for answering both questions, namely from the brain to food choices and
from metabolic profiles (and body composition) to the brain. These brain MR phenotypes include:
MRI sequence Structure and physiology
T1-weighted Volumes, thickness, folding, shape, tissue density
Diffusion tensor imaging Fractional anisotropy, mean diffusivity, track delineation
Magnetization transfer Myelination index
Resting state functional Spontaneous cerebral networks; functional connectivity
Paradigm-based functional Brain response associated with specific stimuli/tasks; functional
connectivity.
B1486 - Phthalates Asthma and Obesity in a Large Longitudinal Cohort - 17/01/2013
Asthma and obesity are two of the most common chronic and disabling conditions of childhood. While both are multifactorial (risk factors include genetics, race, socioeconomic status, diet, and physical activity), most of these risk factors are not amenable to modification or avoidance. However, environmental factors are amenable to change and are plausible contributing factors to both. Given documented rapid increases in both conditions over the past three decades, it is imperative to find ways to reduce both asthma and obesity.
Phthalates, chemicals used to produce a diverse array of consumer products including shampoos, plastic bottles and cosmetics, have been found to interact withperoxisome proliferator-activated receptors that play critical roles in lipid and carbohydrate metabolism.Di-2-ethylhexylphthalate (DEHP) is of particular concern, because mono-(2-ethylhexyl) phthalate,a DEHP metabolite,increases expression of threeperoxisome proliferator-activated receptors(PPARs) which play key roles inlipidand carbohydrate metabolism, providing biological plausibility for a role of DEHP metabolites in childhood obesity and insulin resistance. Additionally, DEHP metabolites induce the release of pro-inflammatory cytokines from lung cells, and activation of PPARs can also modulate immune response.
Investigators have found associations of exposure to plastic wall materials with the development of bronchial obstruction, persistent wheeze, cough, and phlegm in children.Prenatal exposure to butylbenzylphthalate, a high molecular weight (HMW) phthalate used in flooring, has been associated with the development of eczema in one urban longitudinal birth cohort, and a cross-sectional study has associated urinarymono-carboxyoctyl phthalate and mono-carboxynonyl phthalate with asthma.Cross-sectional studies and one longitudinal cohort study have associated lower-molecular weight phthalates with child and adolescent obesity. It is plausible that fetal vulnerability to DEHP is greater, and that this earlier life exposure is more likely to disrupt endocrine processes that maintain dietary balance, leading to obesity. Measurement of phthalates at a single timepoint in pregnancy has moderate sensitivity (56-67%) and high specificity (83-87%) for four phthalate metabolites to estimate exposure tertile over a three-month period, but past studies have been unable to assess a developmental window of vulnerability to phthalate exposure.
The Avon Longitudinal Study of Parents and Children (ALSPAC) is a longitudinal population-based birth cohort study of 14,541 UK mothers enrolled during pregnancy in 1991 and 1992, with data collected at multiple time points during pregnancy and in childhood, through review of hospital records, clinical and laboratory examination, and surveys of parents and children.This population is well-characterized with regard tosociodemographic and behavioral risk factors for obesity, and represents an efficient method to examine ubiquitous environmental chemical exposures as separate risks. We propose to analyze 1500 banked maternal urine samples from each of three trimesters of pregnancy for urinary phthalates, and assess associations with standardized measures of infant, child and adolescent body mass, allergy, and respiratory outcomes.
Aim 1. To examine whether prenatal urinary phthalate metabolites are associated with weight-for-length and Body Mass Index Z-scores in childhood, fat mass, and cardiovascular risks in the school age years and adolescence.
H1. Prenatal urinary phthalates are independently associated with increases in standardized measures of body mass and obesity in childhood, as well as increases in fat mass, adjusting for maternal, sociodemographic and other lifestyle factors.
Aim 2. To examine whether prenatal urinary phthalate metabolites are associated with wheeze, allergy, and asthma phenotype in childhood.
H2a. Prenatal urinary phthalates are associated with increased odds of wheeze in children, adjusting for potential confounding factors.
H2b. Prenatal urinary phthalates are associated with increased odds of allergic phenotype (eczema, rhinitis, or allergy) in children, adjusting for potential confounding factors.
H2b. Prenatal urinary phthalates are associated with asthma phenotype (asthma diagnosis or bronchodilator responsiveness) in children, adjusting for potential confounding factors.
Aim 3. To examine whether prenatal urinary phthalate metabolites are associated with decrements in pulmonary function in the school-age years.
H3. Prenatal urinary phthalates are independently associated with decrements in pulmonary function (forced expiratory volume in the first second (FEV1) and the ratio of FEV1 to forced vital capacity.
B1485 - The influence of breathing disorders on face shape - 17/01/2013
Aims: To explore the effect of breathing disorders including asthma, atopy allergic rhinitis and sleep disordered breathing on face shape at age 15 years and to evaluate the effect of adenoidectomy and tonsillectomy on face shape.
Hypothesis: Breathing disorders, adenoidectomy and tonsillectomy have no effect on face shape in late adolescent.
Background: Respiratory activity may have an influence on craniofacial development and interact with genetic and environmental factors. It has been suggested that certain medical conditions such as asthma have an influence on face shape. Altered mechanics of breathing may influence the development of craniofacial structures and tends to result in significant changes in face shape particularly increased face height and retrusive mandible. However, removal of adenoids and tonsils has been reported to have a significant effect on obstructive breathing and if conducted early will normalize dentofacial morphology. ALSPAC provides the opportunity to explore a longitudinal data set with detailed facial morphology measures in late childhood on a large, representative sample.
Three-dimensional laser surface facial scans were obtained. Differences in the twenty-one reproducible facial landmarks (x, y, z co-ordinates) for the various groups will be evaluated as well as average facial shells will be created for asthmatic, atopic allergic rhinitis, sleep disordered breathing and healthy controls to explore surface differences.
Study design: observational longitudinal cohort study.
Confounding factors: height, weight, BMI, pubertal status.
Freeman K, Bonuck K 2012 Snoring, mouth-breathing, and apnea trajectories in a population-based cohort followed from infancy to 81 months: a cluster analysis. International Journal of Pediatric Otorhinolaryngology 76: 122-30
Jefferson Y 2010 Mouth breathing: adverse effects on facial growth, health, cademics and behavior. General Dentistry 58: 18-25
Wenzel A, Hojensgaard E, Henriksen J M 1985 Craniofacial morphology and head posture in children with asthma and perennial rhinitis. European Journal of Orthodontics 7: 83-92
Zettergren-Wijk L, Forsberg C M, Linder-Aronson S 2006 Changes in dentofacial morphology after adeno-/tonsillectomy in young children with obstructive sleep apnoea--a 5-year follow-up study. European Journal of Orthodontics 28: 319-26.
B1484 - Post natal depression - GWAS meta-anaylsis and allelic risk scores - 17/01/2013
AIMS:
1) To perform a genome-wide association study (GWAS) of postnatal depression (PND)
2) To investigate whether the power to detect PND is improved by incorporating information from nominally significant SNPs in a polygenic risk score
HYPOTHESIS:
1) We hypothesise that common variants contribute to the susceptibility to PND, and that a meta-analysis of GWAS studies can identify regions containing variants relating to this trait.
2) Additionally we hypothesise that our ability to predict PND will be improved by increasing the SNPs included in a genetic risk score, even if they do not achieve genome-wide significance.
VARIABLES:
The outcome variable will be PND case-control status. Women will be classified as cases or controls based on their score in the Edinburgh Postnatal Depression Scale (EPDS) measured 8 weeks postpartum. For the profile scoring, the exposure variable will be the polygenic risk score as calculated from the imputed genotype data available for mothers in ALSPAC. Association results from the Psychiatric Genetics Consortium meta-analysis of bipolar disorder (PGC-BPD) will be used to determine the SNPs to be included in the risk score. Several versions of this risk score will be computed using p-value thresholds of varying stringency (pless than 0.001, 0.01, 0.05, 0.1, 0.5, all SNPs) for inclusion in the score. Ancestry principle components will also be adjusted for.
METHODS:
Mothers enrolled in the ALSPAC cohort who have contributed genetic data to the study, and on whom we have information relating to postnatal depression at 8 weeks postpartum, will be used as the study sample.
We will perform a GWAS using the binary PND variable, and these results will be meta-analysed together those from previous GWAS studies. Polygenic risk scores will be derived using the clumped association results available from the PGC-BPD study. A number of p-value thresholds will be used to determine which of these SNPs to include in the risk score for PND.
Women will be classified as cases or controls based on their score in the Edinburgh Postnatal Depression Scale (EPDS) measured 8 weeks postpartum. A logistic regression model will be constructed, including the risk score and adjusting for both ancestry principle components and season of birth. Each version of the risk score will be incorporated into this model in turn to distinguish between PND cases and controls.
Participants will be excluded from the analysis either if they have a history of BPD or if they experienced a stillbirth or early neonatal death (within 27 days).
B1482 - Using Mendelian Randomisation to clarify the causal relationship between tobacco use and other substances - 03/01/2013
Aims:
To assess the theory of the Gateway Hypothesis, whereby the use of substances such as tobacco and alcohol ('Gateway Drugs') are a risk factor for the use of harder drugs such as cannabis, cocaine and opiods.
Hypotheses:
The hypotheses of this project can be subdivided into three phases:
1. Assessment of the relationship between patterns of smoking and patterns of alcohol consumption in adults. These hypotheses will be tested using the MR technique, where patterns of smoking can relate to either daily cigarettes smoked or short-term smoking cessation.
2. Assessment of the relationship between patterns of smoking and patterns of alcohol consumption in adolescents. This hypotheses will be tested used either standard epidemiological techniques or Mendelian randomisation.
3. Assessment of the relationship between cannabis and other illegal substances. These hypotheses will be tested using hair sample data taken from adolescents.
There is also scope for further analysis linking hypotheses 1 and 2 with that of 3; however this will be assessed at a later date when preliminary results have been obtained.
The following exerpt it taken from the proposal written in application of the PhD studentship for which this research is a part of and will explain all variables being used in the study:
Phase 1 - Relationship between smoking and alcohol in adults
SNP rs1051730 has been shown to be causally related to both heaviness of smoking and smoking cessation in adults. With this in mind, I propose to use the data collected from the mothers in the ALSPAC cohort in order to use MR to assess both the relationship between heaviness of smoking and alcohol consumption and smoking cessation and alcohol consumption. With regards to heaviness of smoking, the exposure variable used will cigarettes smoked per day in the sample of those who smoke, and for smoking cessation, smoking status before and during pregnancy. The outcome variable used for alcohol consumption will focus on heaviness of use.
Once observational epidemiology has been used to initially assess the relationship between tobacco and alcohol, the relationship between genotype and smoking status will be tested using a model with and without interaction terms between the two variables, using a likelihood ratio test. The relationship between genotype and alcohol consumption will also be tested in this manner in order to show whether the genotype is associated with alcohol consumption independently of smoking status. The effect of both heaviness of smoking and smoking cessation on alcohol consumption will be assessed using instrumental variable estimation by performing a two stage least squares under an additive model. This method will first fit the regression of the exposure (smoking) on the instrument (genotype) before the outcome (alcohol) is regressed on the predicted values of continuing smoking, where the estimate of the causal effect will be the coefficient produced. I will also test whether there is a direct effect of genotype on alcohol consumption in non-smokers.
Phase 2 - Relationship between smoking and alcohol in adolescents
To further the work undertaken in phase1 of this research, the relationship between heaviness of smoking and alcohol consumption and adolescents will be assessed using the ALSPAC child based questionnaires and interviews. As a current genetic signal for smoking in adolescence is not known, this work will explore the availability of a polygene risk score for use with MR, or use a standard epidemiological technique of this is unavailable. The exposure and outcome variables used for this section of analysis will be smoking initiation and heaviness of alcohol use.
The first stage of this analysis will be to gather cross sectional studies that have included the rs1051730 across a range of ages in order to assemble a longitudinal map of the effect of this variant, as has been previously done with the FTO gene.
If a genotypic signal is available for use in this analysis then statistical methods for the rest of the phase will be as for phase 1, however if a standard epidemiological method is applied then multinomial logistic regression will be used with smoking as the exposure and alcohol as the outcome with confounding from environmental factors taken into account.
Phase 3 - Relationship between cannabis and illicit drugs in adolescents
Finally, the relationship between cannabis use and the use of other substances as described in the Gateway Hypothesis will be assessed in ALSPAC adolescents using drug information obtained from hair samples. The hair samples can be used in one of two ways: (1) to validate the self report measures taken from the ALSPAC adolescents and (2) to use the information provided by the hair samples itself. With regards to the first option, this method will be used in assessment of cannabis, as hair samples are only able to provide information of regular or heavy use in relation to this drug. The hair sample analysis for other drugs is more sensitive than for cannabis, meaning that this data can both be used on its own and to validate the self report measures, therefore, for this section of the analysis we will use the biological marker as the outcome. As hair samples were not taken from all members of ALSPAC, this section of the study will have a smaller sample size than others; however, the loss of power from the biological marker will be made up for by increased precision and lack of bias. In order to link this with previous phases, cotinine levels of each of the ALSPAC adolescents will also be used to assess the relationship between tobacco and cannabis, ecstasy, cocaine, amphetamines and heroin.
Standard epidemiological statistical methods will be used here, with cannabis use from self reported questions (validated by hair sample data) as the exposure and the use of other illicit substances as shown by hair samples as the outcome.
Analysis:
All analysis will be undertaken by University of Bristol staff (Michelle Taylor) with Glyn Lewis, Matthew Hickman and Marcus Munafo acting as PhD supervisors.
B1480 - Investigation of fascin-2 and associated components of hair cell stereocilia in relation to ALSPAC hearing data - 20/12/2012
Aim.
To investigate whether genetic polymorphisms/mutations in the FSCN2 gene, or in genes encoding proteins known to be present in hair cell stereocilia and associated physically or functionally with fascin-2 protein, are correlated with early-onset hearing loss within the ALSPAC study population.
Hypotheses
1. The human population includes mutations/polymorphisms in FSCN2 that are associated with early-onset hearing loss.
2. The human population includes mutations/polymorphisms in genes encoding proteins that are linked physically or functionally with FSCN2 for assembly and mechanotransduction by stereocilia of hair cells, that are associated with early onset hearing loss.
Rationale
The hair cells of the inner ear have essential roles in mechanotransduction of sound waves into intracellular signals that result in nerve cell impulses. Crucial to mechanotransduction are the hair bundles on the apical surfaces of hair cells. These are made up from many actin-rich stereocilia, fine projections of the cell surface that are made rigid by a central bundle of cross-linked actin filaments. It has recently become appreciated that major components of a protein complex (tip complex) located at the tip of stereocilia are important for gating ion channels in response to mechanical movements of the stereocilia, and that function-perturbing mutants in multiple members of the the tip complex lead to deaf/blindness syndromes such as Usher syndrome (Bonnet and El-Amraoui, 2012). An example of one of these proteins is cadherin-23. Mis-sense mutations in cadherin-23 are also associated with non-syndromic deafness (Schultz et al., 2011). The range of phenotypes associated with cadherin 23 mutations suggest that additional interactions between components of the stereocilia and potential genetic associations with early-onset hearing loss remain to be discovered.
Fascin-2 is an actin-crosslinking protein that until recently was considered to be located specifically in photoreceptor cells of the retina (Lin-Jones and Burnside, 2007; Hashimoto et al., 2011). New data demonstrate that fascin-2 is a component of inner and cochlear hair cell stereocilia in humans, mice and fish (Shin et al., 2010, Chou et al., 2011). Furthermore, studies of very early onset hearing loss (phenotype: increased 16kHz hearing threshold by 5 weeks of age) in the inbred DBA/2J strain of mice have identified a casual role for a mutant variant of fascin-2, fascin-2R109H, in synergy with a mutation in cadherin-23 (Johnson et al., 2008, Shin et al., 2010). This phenotype of these mice can be rescued by transgenesis of wild-type fascin-2 (Shin et al., 2010).
Many SNPs, including multiple non-synonymous coding single nucleotide polymorphisms (SNPs) have been identified in human FSCN2 (NCBI SNP database). Thus a major goal of this pilot study is to identify if any polymorphism in human FSCN2 is associated with early onset hearing loss. We will also examine possible contributions of fascin-2 associated proteins. The large ALSPAC cohort of children who have had hearing measurements taken, including the very sensitive measurements of 16kHz audiometry and otoacoustic emissions and tympanometry variables for middle ear function, can enable us to address this question.
Design
1. The first hypothesis we will test is that SNPs in human FSCN2 and CAD23 are associated with early onset hearing loss in humans, with emphasis on results of extra-high frequency audiometry and otoacoustic emissions, but also including tympanometry variables for middle ear functions.
2. Secondly, we will examine for possible associations of SNPs in genes that encode other proteins of the stereocilia that associate physically or functionally with fascin-2 or cadherin-23. These will include:
-protocadherin-15, that acts with cadherin-23 in the tip links of the tip complex
-other actin crosslinking proteins that are present within stereocilia, as determined by proteomic analyses of isolated stereocilia (plastin-1 as the second most abundant cross-linker, also espin, plastin-2, plastin-3, fascin-1, espin-like protein and xin-related protein 2) (Shin et al., 2010)
-the ion channel HCN2, reported in a biochemical study to associate physically with fascin-2 in cochlear hair cells (Ramakrishnan et al., 2012)
-the human orthologues of six gene products from a candidate interval on mouse chromosome 18 that appears to account for broader spectrum, low frequency hearing losses that occur in the DBA/2J mouse strain by 2-3 months of age (Nagtegaal et al., 2012). The syntenic region of the human genome is 18q21.1 and the six genes are all conserved in this region (NCBI Genes database); MYO5B, ACAA2, c18orf32, DYM, SMAD7 and CTIF
Associations will be examined for each individual gene, and we will also examine if any coding or other SNPs identified occur in the same individual as any SNPs identified in fascin-2 or cadherin-23.
3. To make these analyses, we request access to all SNPs examined within the children in the ALSPAC hearing data studies for the following genes:
Priority 1:
FSCN2
CDH23
HCN2
PCDH15
PLS1
Priority 2:
ESPNL
FSCN1
PLS2
PLS3
XIRP2
MYO5B
ACAA2
c18orf32
DYM
SMAD7
CTIF
References
Bonnet C, El-Amraoui A. (2012) Curr Opin Neurol. 25:42-9.
Schultz, JM et al. (2011) J Med Genet. 48:767-75
Lin-Jones J, Burnside B. (2007) Invest Ophthalmol Vis Sci. 48:1380-8.
Chou SW et al. (2011) PLoS One. 2011;6(4):e14807
Johnson KR et al. (2008) Genomics 92: 291-225.
Shin JB et al., (2010) J. Neurosci 30: 9683-94.
Hashimoto, Y, Kim DJ and Adams, JC (2011). J. Pathology 224:281-300.
Ramakrishnan NA et al (2012) J Biol Chem. 287:37628-46.
Nagtegaal AP et al. (2012) Genes Brain Behav. 2012 Sep 18. doi: 10.1111/j.1601-183X.2012.00845.x.
B1483 - DXA project - 20/12/2012
There is a DPhil computer science student from Oxford University on secondment to a commercial company who may be able to automate our novel method for measuring spinal curvature from total body DXA scans. We developed this method as part of the BSRF-funded ALSPAC project B727, and the method has been submitted for publishing and has been presented in abstract form so is in the public domain.
B1481 - Explaining the association between socioeconomic status and cognitive growth in childhood - 20/12/2012
Differences in cognitive ability determine developmental trajectories across the lifespan, affecting socioeconomic, psychological, and health outcomes. Children from disadvantaged socioeconomic status (SES) families show lower cognitive ability and reduced cognitive growth (i.e. ability gains over time) compared to their high SES peers. However, the mechanisms that underlie the SES-growth association are not fully understood. The proposed research investigates how and to what extent 3 aspects of children's lives - (1) dietary patterns, (2) preschool experiences, and (3) activity engagements - account for SES-related growth differences.
(1) Dietary Patterns. Diets directly impact cognitive function because they are the main source for micronutrients (e.g. vitamins and antioxidants) that enable enzymatic reactions and neurotransmitter synthesis in the brain. However, several diet characteristics have rarely been addressed in research, for example meal types (e.g. snack), times (e.g. breakfast), and preparation method (e.g. prefabricated). Likewise, no previous research has looked at the influence that children's drinking habits (e.g. milk, sugary drinks) may have for cognitive development.
(2) School Experiences. SES-related differences in primary school performance tend to increase or at least persist over time, with disadvantaged children continuing at lower quality secondary schools and obtaining fewer educational qualifications. Preschool attendance may reduce this achievement gap because it provides disadvantaged children with better learning environments than they receive otherwise. The benefits of preschool have not been studied in Britain, where several preschool types exist (e.g. nursery or childcare centre).
(3) Activity Engagement: The availability of activity engagement opportunities, for example music lessons or sport clubs, leads to better primary school achievement16 and is thought to benefit children's cognitive growth. High SES families encourage more frequent and diverse activity participation than do low SES families in the United States16 but these findings have yet to be replicated in Britain. Also, it is unclear if it is the quantity, quality or type of activity engagement that make the difference for children's cognitive development.
Methods: Data will be synchronized across 4 longitudinal, British cohort studies, including the British Cohort Study 1970 (BCS), the Growing Up in Scotland study (GUS), the Millennium Cohort Study (MCS), and the Avon Longitudinal Study of Parents and Children (ALSPAC). The cohorts comprise of large samples, representative of the British population, and span generations from 1970 to 2005 (Table 1). Thus, they allow for (1) exploring and (2) replicating effects of dietary patterns, preschool experiences and activity engagements (herein variable complexes) on cognitive growth, as well as for (3) analyzing possible generation-specific effects. In each cohort study, participants were assessed at least 3 times on cognitive ability, enabling (4) latent growth curve modelling of ability trajectories. They also completed several, mostly repeated measures for all variable complexes, permitting (5) in-depth investigations of their characteristics and continuity over time.