Proposal summaries
B2114 - The role of environment and executive function in the social patterning of health-related behaviours - 28/11/2013
Aim:
To assess the extent to which effects of environment on health-related behaviour are attenuated by executive function. In particular, to assess whether high executive function (in particular, response inhibition) negates the impact of exposure to environments that are more likely to promote less healthy choices.
Four health-related behaviours - diet, physical activity, smoking and alcohol consumption - will be explored, using measures taken between the ages of 10 and 15.
B2113 - Maternal pre-eclampsia and bone mineral density of the adult offspring - 28/11/2013
Aim: To assess bone mineral density (BMD) and bone mineral content (BMC) of offspring born from a preeclamptic pregnancy.
Hypothesis: We hypothesise, based on the limited available evidence on this topic, that offspring who were 'exposed' to preeclampsia will have a higher bone mineral density than those who were not exposed.
Exposure variables: Preeclampsia and gestational hypertension
Outcome variables: Bone mineral density and content (as measured by DXA and pQCT)
Confounding/mediator variables: SES, maternal smoking, maternal BMI, maternal ethnicity, maternal age, maternal alcohol consumption, maternal physical activity, maternal vitamin D & Calcium intake, offspring weight/height, offspring total body fat mass/lean mass, CTX, Insulin, parity, birthweight, gestational age.
Summary:
Preeclampsia is the leading cause of perinatal and maternal mortality and morbiditiy and effects around 3% of all pregnancies worldwide. The only current cure is the delivery of the placenta and thus often results in an iatrogenic preterm delivery. The fetus has the greatest demand for calcium during the third trimester, so consequently a preterm delivery can result in insufficient bone mineralization (Weiler et al. 2002). A recent study has suggested that in pregnancies compliated by preeclampsia additional mechanisms may exist which result in protective effects on long term bone health in the offspring, both in preterm and term births (Miettola et al. 2013). However, this study was limited by a small sample size and no maternal lifestyle information. This study warrants replicating with larger numbers and more comprehensive confounder adjustment.
We hope to use the ALSPAC data to explore the association between preeclampsia (and gestational hypertension) in term and preterm births with bone parameters from recently collected DXA and pQCT scan data.
References
Miettola S, Hovi P, Andersson S et al. Maternal preeclampsia and bone mineral density of the adult offspring. Obstetrics. Am J Obstet Gynecol 2013;20:443.e1-10
Weiler HA, Yuen CK, Seshia MM. Growth and bone mineralization of young adults weighing less than 1500g at birth. Early Hum Dev. 2001; 67: 101-12
B2112 - Exploring the fetal insulin hypothesis in ALSPAC - 28/11/2013
Aims
1) To investigate associations between offspring birthweight and offspring fasting glucose/insulin at birth, age 8.5 years, 15 years and 17 years.
2) To investigate associations between offspring birthweight allele score (known variants and genome wide prediction), birthweight and offspring fasting glucose/insulin
3) To investigate associations between offspring birthweight allele score (known variants and genome wide prediction), birthweight and offspring fasting glucose/insulin - with adjustment for maternal genotype to remove negative confounding.
4) To investigate associations between maternal birthweight and maternal fasting glucose/insulin in FOM1 clinic
5) To investigate associations between maternal birthweight allele score, birthweight and maternal fasting glucose/insulin
6) To investigate associations between offspring glucose/insulin allele score and offspring birthweight
7) To investigate the timing of effect of individual glucose/insulin variants
8) Bivariate GCTA of offspring birthweight and offspring glucose/insulin.
B2111 - Identification of patterns of methylation associated with age at menarche - 28/11/2013
Aims:
1) To identify differentially methylated regions (DMRs) that are associated with age at menarche. This will be tested in both ALSPAC mothers using recalled age at menarche, and in young participants using DNA methylation measured at 15-17 years and age at menarche reported at the 21/22 year visit.
2) To explore whether identified DMRs are already apparent at birth, or if they are established over the course of childhood and adolescence.
3) To understand how common genetic variation underlies differences in methylation relating to menarche timing.
B2110 - DIETARY PATTERNS IN PREGNANCY AND OFFSPRING GROWTH OUTCOMES A MULTI-COUNTRY ANALYSIS OF BIRTH COHORTS - 28/11/2013
Aims:
1) Characterise dietary patterns in pregnancy using data from English and Danish birth cohorts.
2) Compare dietary patterns between pregnant women living in England and Denmark.
3) Examine the relationship between maternal dietary patterns in pregnancy and size at birth.
4) Examine the relationship between maternal dietary patterns in pregnancy and offspring growth at year 7, 1 and 6 months.
5) Assess the relationship of change in maternal dietary patterns and childhood adiposity measures.
B2109 - Depression at 17 THEME 3 CONTINUITIES AND DISCONTINUITIES - Depression and Fatigue part of previous Wellcome Trust application - 14/11/2013
Adolescent depression is highly comorbid with other disorders, including persistent fatigue or chronic fatigue syndrome. Teenage years are a key period of change and provide an opportunity to look at the continuities and discontinuities of psychopathology from a developmental perspective. Heterotypic continuities could arrise in a variety of ways: because the two disorders share common or correlated risk factors (genetic or environmental); because one disorder contributes, directly or indirectly, to risk for the other; or because the joint pattern constitutes a distinct diagnosis entity. ALSPAC has repeated measures of psychopathology. These analyses will investigate the relationship between depression and fatigue, the direction of causality and potential explanatory variables such as physical activity levels and early stressful life events. This study will examine the role of both parental teenagers, including maternal depression and early childhood depression. The study will further examine a range of hypotheses about premorbid risk markers for persistent fatigue including the role of personality, childhood adversity (material and emotional), socioeconomic and physical activity factors.
B2108 - Quantifying the harms and costs of passive drinking in the UK - 14/11/2013
The overall aim of this proposal is to address the MRC, ESRC and Alcohol Research UK: Call for Alcohol Misuse Research theme "Quantifying the harms to non-drinkers". Such secondary harms or harms from passive drinking refer to the indirect harm (e.g. cognitive, developmental, psychological, domestic violence, crime, injury, and loss of productivity) to others (e.g. foetus, child/ren, partner, and society) resulting from someone else's drinking. Therefore, the overall aim of this project is to quantify the social, psychological and economic costs to families from either one or both parents' alcohol consumption.
To determine the indirect harms of parental drinking on the child, the project will:
1) longitudinally quantify the indirect harms across the child's lifespan (from foetus to aged ) from parental drinking on neurodevelopment and educational attainment, psychological well-being and substance use for the child
2) consider whether there are sex differences in indirect harms experienced
3) determine the harm resulting from different patterns of drinking over time, and whether changes in alcohol consumption/patterns of drinking result in changes in the associated harm
4) determine protective factors (resilience)
5) determine the combined harm of alcohol and domestic violence
6) determine the associated economic costs of parental drinking on the child by examining service utilization
To determine the indirect harms of partner drinking on the other partner, the project will:
1) longitudinally quantify the indirect harms across the partner relationship on psychological well-being and substance use
2) consider whether there are differences in indirect harms experienced, depending on the sex and sexuality of the partners
3) determine the harm resulting from different patterns of drinking over time, and whether changes in alcohol consumption/patterns of drinking result in changes in the associated harm
4) determine protective factors (resilience)
5) determine the combined harm of alcohol and domestic violence
6) determine the associated economic costs of parental drinking on the child by examining service utilization.
B2106 - Predicting oral language development to middle childhood and its implications for later achievement and wellbeing - 07/11/2013
Aim
To evaluate the potential of early life biological, parental and socio-demographic factors to predict natural language skills, and the extent to which these skills have an impact on school outcomes in (late) adolescence.
B2105 - Does eye colour change between 3 months and 5 years of age - 07/11/2013
Perceived eye, or iris colour varies as a continuous from blue to brown/black and is determined by the melanin pigment, quantity, packaging and type produced in the melanocytes of the anterior border layer and stroma of the iris. Brown eye colour is the result of abundant melanin pigment content and a higher ratio of pheomelanin to eumelanin. The number of melanocytes is not a determining factor. Blue eye colour results from the scattering of most blue wavelengths to the surface by collagen fibrils, in the presence of minimal melanin pigment. (1)
Once thought to be an example of simple Mendelian inheritance where brown colour was dominant and blue colour recessive, iris colour is now known to be a polygenic phenotypic characteristic with a myriad of possible combinations of parent and child eye colours. For example, two blue-eyed parents may produce a brown, green-hazel, or blue-eyed child. (1)
It is accepted that infant eye colour at birth does not necessarily correlate with resultant childhood and indeed adult eye colour, although it is found to be roughly stable by six years of age. (2) There is much anecdotal evidence of eye colour changing from one extreme of the spectrum to other in the first year of life. One study investigated this, finding a decrease in the number of light irises. (3) Congenital conditions affecting iris colour include albinism, in which the iris may be red or violet and heterochromia in which the two irises are different colours. (4)
The main contributory genes for iris colour are HERC2 (hect domain and RCC1-like domain-containing protein 2) and OCA2 (oculocutaneous albinism II) on chromosome 15, which account for around 74% of variation. Depression of OCA2 transcription results in a blue-eyed individual. A further 14 genes have been identified as contributing, however many of these play a greater role in determining skin and hair colour. Given the complex inheritance model of eye colour it is therefore difficult to predict final outcome based on the eye colour of the parents alone. (1, 4)
Many parents in the neonatal unit, ophthalmology and paediatric clinics enquire about changes in their child's eye colour, however literature is limited on the topic and the timing and nature of the change in eye colour in healthy neonates and young children is not well described.
ALSPAC is a vital resource in the study of child development and has contributed to much influential work on the topic of child development and paediatric disease. Simple data were collected regarding iris colour at various stages in the first 2 years of life and later at age 5. To our knowledge, this data has not been analysed to look at change in eye colour over time.
Understanding the nature of eye colour change in young children will broaden the knowledge base of both clinicians and patients or relatives and may aid in the understanding of iris development, which is important as it appears to be closely linked with brain development. (1) To date eye colour has not been found to be associated with visual acuity, but has been linked with reaction time, shyness, noise-induced hearing loss and ability to overcome seasonal affective disorder. (4) The individual iris is so unique in terms of colour, pattern and features that it can be reliably used in identification in place of a passport. (5)
The study team would be interested to investigate whether there was a change in iris colour between the age of 3 months and 5 years; at what point these changes occur and what form these changes take.
The study team are experienced in working with ALSPAC data and have paediatric ophthalmological clinical experience. Dr Creavin is an academic ophthalmology trainee whose research interest is ophthalmic epidemiology.
References
1. Sturm RA & Larsson M. Genetics of human iris colour and patterns. Pigment Cell Melanoma Res. 22;544-62
2. Bito LZ, Matheny A, Cruickshanks KJ, Nondahl DM, Carino OB. Eye color changes past early childhood: The Louisville Twin Study. Arch Ophthalmol 1997;115:659-63
3. Matheny AP and Brown Dolan A. Changes in eye colour during early childhood: sex and genetic differences. Annals of Human Biology 1975;2(2):191-6
4. White D and Rabago-Smith M. Genotype-phenotype associations and human eye color. Journal of Human Genetics 2011;56:5-7
5.Dougman J. Probing the uniqueness and randomness of IrisCodes: results from 200 billion iris pari comparisons. Proc. IEEE 2006;94:1927-35
B2103 - Communication difficulties peer victimization and mental health - 31/10/2013
AIMS
The main aim of the study is to explore the relationships among speech, language and communication needs, peer victimisation during childhood and adolescence, and mental health across the lifespan. The research questions are:
* Are levels of reported peer victimization across different SLCN groups (such as stuttering, other speech difficulties, and developmental language difficulties) significantly greater than for cohort members without such needs? Where possible we will take into account data from cohort members with non-SLCN disabilities.
* How do measures of mental health and emotional well-being in these groups compare across the lifespan?
* To what extent does peer victimization contributes to mental health in these groups across the lifespan?
* Does SLCN remain a risk when other factors are controlled for?
* Have levels of reported peer victimization in SLCN groups changed in response to initiatives to tackle bullying?
A further aim of the study is to develop a framework for future collaborative work involving the co-applicants and in the longer term other birth cohort researchers in the international academic community.
We plan to use data from the National Child Development Study and the Millenium Cohort Study as well as ALSPAC, because we wish to explore the impact of SLCN across the lifespan. However, ALSPAC offers the richest source of information in many respects. First, parental report of stuttering and developmental speech difficulties is verified by a trained researcher, which is not the case for any other British birth cohort study to date; second, parents, teachers and the cohort members themselves provide report of peer victimization; third, it provides a wide range of mental health measures.
Mental health and peer victimisation will both be considered as outcomes in different models. We will consider a variety of models as appropriate for the different outcome variables, including different types of multivariable regression such as binary and multinomial logistic regression and hierarchical linear regression. We would also conduct path analysis to examine direct pathways between SLCN and mental health, as well as indirect pathways via peer victimization over and above the effects of individual and socio-demographic factors and life events.
HYPOTHESES
Based on prior research, we hypothesise that cohort members with SLCN will be more likely to experience bullying and mental health problems than controls without these conditions. As suggested above, we expect that we will find direct pathways between SLCN and mental health, as well as indirect pathways via peer victimization, over and above the effects of individual and socio-demographic factors and life events.
B2102 - Novel Methods to Study Diet andObesity Compared to Genetics and Other Lifestyle Factors Agent-Based Modelling - 31/10/2013
Our specific aims are:
1. Derive empirical lifestyle patterns using cluster and/or factor analysis in a cohort of British
children.
2. Develop a mathematical model to simulate the association between derived dietary patterns and
other lifestyle factors (breakfast consumption, family dinner, breastfeeding history) and demographic
covariates (age, sex, socioeconomic status) with body fat using agent-based modeling (ABM).
3. Test the model using data collected from children aged 10 to 13 years over 4 time points (2004,
2006, 2008) to determine which variables are most strongly associated with changes in body fat over time
as children transition into adolescence.
4. Examine whether and how associations between dietary patterns, lifestyle factors, and obesity
shift when genetic data are included in the model (subset analysis*).
5. Reproduce aims 1 through 3 in a large prospective cohort of American children aged 10 to 13 y.
B2101 - Shared genetic determination of refractive error with age and ethnicity - 31/10/2013
Aims: To assess the genetic correlation between refractive error in (a) European vs. East Asian adults, (b) European children vs. European adults, and (c) European children vs. East Asian adults.
Hypothesis: The above genetic correlations are high (i.e. greater than 0.5), which implies that the higher prevalence of myopia in East Asians compared to Europeans is not simply genetic in origin.
Variables: Refractive error will be based on autorefractor measurement. Within each of the 3 subject cohorts, refractive error will be rank-transformed to a give a Normal Score, using Blom's method. (Our ultimate aim is to compare the refractive error of individuals of different ethnicity/age after accounting for absolute differences due to the environment in which they grew up. Also, heritability analysis is not robust against departures from normal trait distributions).
Analysis method: The degree of genetic relatedness between all pairs of individuals in 2 cohorts combined will be calculated using the GCTA program (for the 3 sets of 2 cohorts listed as a, b and c in the Aims above). Bivariate variance components analysis will be carried out using GCTA, as described [4].
Preliminary work: We have used bivariate GCTA in ALSPAC YPs to show that the shared SNP-heritability across childhood is high (greater than 0.7) (unpublished results). Too few ALSPAC mothers have refractive error information for estimation of its SNP-heritability (unpublished results).
Precautions to ensure the identity of participants cannot be revealed: Under certain circumstances, access to high-density genotype data is sufficient to reveal the identity (e.g. surname) of an individual [5]. We propose the following scheme to make such identification impossible (given current knowledge):
a) Create a list of the SNPs present on the genotyping platform used for each of the 3 cohorts (ALSPAC, Rotterdam, Nagahama).
b) Identify SNPs genotyped in all 3 cohorts. Sort the list by chromosome, then by genomic position.
c) From the list generated in step (b) randomly delete 5% of SNPs.
d) From the list generated in step (c) randomly re-order the SNPs within chromosomes (i.e. shuffle the order of SNPs on chromosome 1, then shuffle the order of SNPs on chromosome 2, etc.).
e). Send the list generated in step (d) to each site (ALSPAC, Rotterdam, Nagahama).
f) At each site, an analyst links refractive error and genotype data. The order of subjects in the data file is then randomised, and the unique subject identifier (e.g. "ALN") is replaced with a new identifier based on file order, e.g. ALSPAC1, ALSPAC2, ALSPAC3, ... ALSPAC9999.
g) For the data file generated in step (f), the analyst at each site sorts the order of SNPs so that it matches the order of SNPs in the circulated list (e). SNPs missing from the list are deleted.
g) For the data file generated in step (g), the analyst at each site replaces the SNP name with a new name based on chromosome and file order, e.g. chr1snp1, chr1snp2, chr1,snp3, ...chr22snp3455.
h) The analyst at each site sends the file generated in step (g) to a central site for GCTA analysis.
NB. Without knowing the order of SNPs in the circulated list (e), the order of SNPs in the data files (h) is unknown, and thus cannot be used to infer identity....yet it can be used to calculate relatedness of subjects whose SNPs are sorted in the same order.
References
1. Pan CW, Ramamurthy D, Saw SM (2012) Worldwide prevalence and risk factors for myopia. Ophthalmic and Physiological Optics 32: 3-16.
2. Morgan IG, Ohno-Matsui K, Saw SM (2012) Myopia. Lancet 379: 1739-1748.
3. Verhoeven VJM, Hysi PG, Wojciechowski R, Fan Q, Guggenheim JA, et al. (2013) Genome-wide meta-analyses of multiancestry cohorts identify multiple new susceptibility loci for refractive error and myopia. Nature Genetics 45: 314-318.
4. de Candia Teresa R, Lee SH, Yang J, Browning Brian L, Gejman Pablo V, et al. (2013) Additive genetic variation in schizophrenia risk is shared by populations of African and European descent. American Journal of Human Genetics 93: 463-470.
5. Gymrek M, McGuire AL, Golan D, Halperin E, Erlich Y (2013) Identifying personal genomes by surname inference. Science 339: 321-324.
B2100 - Longitudinal genome-wide association study GWAS for refractive error - 31/10/2013
Overview.
We will request funding to employ a post-doctoral RA to carry out these analyses. The RA will be based at the PI's institution (Hong Kong Polytechnic University) and supervised by the PI and co-I's.
We seek approval from the ALSPAC Executive Committee to run the proposed analyses on the blue crystal computing cluster. The analyses are very computationally intensive (fitting a mixed model for each of millions of SNPs). We will include in the grant any funds specified by the Executive Committee required to cover this computing time as well as other appropriate costs incurred by ALSPAC. (NB. The longitudinal GWAS for BMI, which examined ~3 million SNPs, required approximately 8648 days of computing time to complete).
Aim.
To identify genetic variants influencing refractive error trajectory (either in the general population or in subjects classified as spending a high/low amount of time reading or outdoors).
Methods.
The refractive error and genotyping data has already undergone QC filtering for use in a closely related project that has received Exec approval: B1352 'CREAM (Consortium for Refractive Error and Myopia) 1000-genomes meta-analysis'.
Preliminary analysis of refractive error trajectories using linear mixed models are very promising (for example, as part of the above amendment work, we have been able to identify known myopia SNPs that (a) act early in childhood and retain stable effects during childhood, (b) exert a progressively increasing effect during childhood, and (c) interact with time spent outdoors or time spent reading.
However, fitting the model for each SNP currently takes ~1 minute. We propose to test simplified models and examine the robustness of fixed effects estimates, in order to minimse the time required to fit each model. Once optimised, we will conduct three longitudinal GWAS analyses for three models:
Model 1. Test for SNP main effects and 'SNP x age' interaction.
Model 2. As model 1, plus a test for 'SNP x nearwork' interaction.
Model 3. As model 1, plus a test for 'SNP x time outdoors' interaction.
We will aim to examine the ~4.5 million very high confidence imputed (quality metric RSQR greater than 0.8) very common variants (minor allele frequency, MAF >=0.1) SNPs in the 1000G-imputed dataset.
B2099 - Associations of hypertensive disorders of pregnancy with adolescent offspring cardiac function and structure - 31/10/2013
AIMS
* To investigate the associations between exposure to HDP and blood pressure trajectories during pregnancy with offspring cardiac structure and function.
HYPOTHESIS
We hypothesize that the cardiac function and left ventricular geometry in adolescents exposed to HDP is less favorable compared to offspring of normotensive mothers.
METHODS
We will use data from ALSPAC to investigate the association between preeclampsia and offspring cardiac structure and function. Of the 5,217 adolescent children investigated, 2,000 conceded to echocardiography.
The echocardiograph examination
Echocardiography was performed using a HDI 5000 ultrasound machine (Phillips) equipped with a P4-2 Phased Array ultrasound transducer using a standard examination protocol. All measurements were made according to American Society of Echocardiography (ASE) guidelines.
Definition of exposure variables
* Being born to a mother with preeclampsia or GH.
* Blood pressure trajectories during pregnancy
Confounding/mediating variables (available data)
Sex of offspring, birth weight, length of pregnancy, the adolescent's own blood pressure, maternal pre pregnancy blood pressure, maternal pre pregnancy BMI, maternal and offspring smoking and physical activity, age and parity of mother, height in adolescence, weight/fat mass in adolescence, occupational social class, maternal diabetes in pregnancy.
Outcome variables
Cardiac structure
* Left ventricular size indexed to height (LVMI)
* Left ventricular geometry (based on LVMI and left ventricular relative wall thickness)
Cardiac function
* Left ventricular systolic function (Ejection fraction etc.)
* Left ventricular diastolic function (E/A, E'/A')
Statistical analyses
All statistical modeling will be made in SAS (SAS Institute inc., Cary, NC, USA). We will use multivariate linear models to estimate the associations and adjust for potential confounders and mediators. A previous study based on the follow-up clinic at 17 years reported the prevalence of exposure to preeclampsia and GH in utero to be 2% and 15%, respectively.5 Assuming a similar prevalence in the subgroup examined by echocardiography, this would result in 40 study participants having been exposed to preeclampsia and 300 having been exposed to GH.
B2098 - Genome-wide analysis of association between genotype and methylation - 24/10/2013
Outline
DNA methylation data has recently been generated from ALSPAC samples at multiple timepoints using the Illumina HM450k DNA methylation array. Comprehensive genotype data also exists for the same samples. This project will integrate these two large-scale datasets to investigate the role of genetic variation in DNA methylation.
Aims
1. Genome-wide association analysis of methylation in children and mothers to identify both cis and trans associations between genetic variants and methylation
2. Analysis of change in genotype/methylation associations through the lifecourse
3. Comparison of genotype/methylation associations to published genotype/phenotype and genotype/outcome associations
4. Comparison of genotype/methylation associations to genotype/expression associations (published and within ALSPAC)
5. Comparison of genotype/methylation assocations to ENCODE regulatory regions
6. Analysis of proportion of methylation variance that is influenced by genotype (cis/trans) and non-genetic factors.
9. Single point genotype-stage interaction analysis
10. Genetic and non-genetic correlation over space and developmental stages.
B2096 - Relationship between nicotine metabolism rate as determined by CYP2A6 genotype and smoking patterns in adolescence - 24/10/2013
Aims:
1. Examine patterns of cigarette use across ages 13 to 21 years and identify sources of heterogeneity in smoking phenotypes
2. Identify relationships between CYP2A6 nicotine metabolism group and smoking patterns across ages 13 to 21 years
3. Compare the relative influence of CYP2A6 nicotine metabolism group on smoking initiation vs. smoking escalation across ages 13 to 21 years
Hypotheses:
Relative to normal CYP2A6 metabolizers, reduced CYP2A6 metabolizers will:
1. Escalate cigarette consumption faster in earlier adolescence
2. Escalate cigarette consumption slower in later adolescence/young adulthood
3. Have higher levels of cigarette consumption at earlier stages in adolescence
4. Have lower levels of cigarette consumption at later stages in adolescence/young adulthood
We also predict that there will be no association between CYP2A6 metabolism group and smoking initiation.
B2095 - Sex specific differences in the placental syndrome and markers of placentation - 24/10/2013
AIMS
We aim to
1) Investigate sex-specific differences in the occurrence of early onset and late onset pre-eclampsia, SGA and preterm birth
Where studies currently have relevant data and/or in the future we also plan to add to this aim with the following two aims. However, this request is currently ONLY for aim 1 above.
2) Explore sex-specific differences in markers of placentation in maternal blood
3) Undertake (epi)genetic studies assessing methylation / imprinting differences.
B2093 - Gender nonconformity sexual orientation and distress - 24/10/2013
The present study aims to address these previous research gaps and limitations by examining associations between prospective measures of childhood gender non-conformity and subsequent negative social interactions, poorer self-image, and distress. A wider range of forms of negative interpersonal reactions and types of distress will be examined than in previous studies.
B2097 - Analysis of correlation patterns in DNA methylation through the lifecourse - 23/10/2013
Outline
DNA methylation data has recently been generated from ALSPAC samples at multiple timepoints using the Illumina HM450k DNA methylation array. Genome-wide correlation analysis is now possible to (a) identify potential coregulation and (b) to inform dimension reduction strategies to data analysis. It is also possible to explore the correlation between timepoints to determine which CpG sites in the genome are functionally stable, and which change over time.
Aims
1. Windowed analysis across the genome to generate "methylation correlation maps" at all time-points
2. Genome-wide analysis of correlation between variable loci to identify networks of genes that appear to be coregulated
3. Analysis of correlation of methylation between timepoints at each site in the HM450k to identify stable and labile CpGs.
B2091 - An investigation into the confounding associated with genome-wide score and methods to minimise its impact - 10/10/2013
AIM
We plan to investigate the reasons for confounding of the genome-wide genetic BMI score and its extent using in ALSPAC and examine various means to minimise its impact any subsequent analysis.
Hypothesis
In an ongoing study of ALSPAC participants (Deanfield London Study) our recruitment is based on extremes of BMI percentages as determined by BMI associated common genetic variants. We used a genome-wide score that enabled us to reach a difference of 3 BMI units between the top and bottom percentages of BMI categories.
Preliminary data analysis on 200 subjects in the current study suggests that while the genetic score has as expected separated the population into two distinct BMI categories with mean difference of 3 BMI unit difference. However, the score itself is not totally unconfounded from environmental factors as was predicted at the start. This is likely is due to the use of a genome-wide score, instead of using a score from a limited set of strongly associated SNPs, that appears to have re-introduced confounding to some extent, although the reason behind these unexpected properties is unclear.
Methods
We will to explore this confounding and aim to answer the following questions.
a) What is the extent of the confounding of the genetic score i.e. what are the main confounders?
b) What is the main driver of this confounding: Genuine or horizontal pleiotropy? Vertical pleiotropy with SNPs acting on intermediate pathways to BMI or population stratification?
c) Can this confounding be controlled by adjusting for the main confounders?
d) Is this confounding less than what is expected by generation of groups by BMI alone? If yes, by how much?
e) In light of the above what are the limitations of our selected groups with differential BMI for the purpose of using them to study the main programme question i.e. adiposity-cardiovascular disease associations.
To answer the above questions we would require extensive data on confounders, the genetic score groups, obesity profiles and cardiovascular risk factors. In brief the following analysis will be performed.
Using standard regression model we shall investigate if the BMI genetic score is associated with an extensive list of confounders. We shall use adjustment methods for various confounders to investigate the the drivers of confounding and if their effect can be minimized. Intermediate pathway confounders will be exploited to investigate the presence or absence of vertical plieotropy.
Employing PCA (principle component analysis) we shall also investigate if there is a structure in the SNP sets used in the score and can PCA adjustment remove the confounding effect.
Using regression modelling we shall quantify extent confounding using conventional BMI and compare and contrast with the ones revealed from the genetic score analysis.
Using dummy data which shall be used to study the main questions from the ongoing study (Deanfield London Study) i.e. visceral adiposity-cardiovascular disease markers association, this way we can inquire into any limitation of adjustment for these confounding on any subsequent analysis on the real data when its available after study completion.