23andMe is currently carrying out SNP genotyping of the ALSPAC cohort across approximately 580,000 common SNPs with genome-wide coverage. We are using a custom version of the Illumina HumanHap 550+ platform with additional content selected by 23andMe. The additional content includes coverage of known common and rare SNP associations; better coverage of drug metabolizing genes; and dense coverage of the Y chromosome, mitochondrial DNA, and the MHC region. The 23andMe customer cohort and the ALSPAC cohort have substantial differences in demographics, and phenotypes have been assessed very differently. Essentially all 23andMe customer phenotype data is self reported through web based surveys, and unlike ALSPAC, the 23andMe cohort is far from being a cross section of an identifiable population. As a result, we think it is most appropriate to model SNP associations separately in the two cohorts and then combine association test results in a meta-analysis. For this initial data access proposal, we have enumerated a subset of the ALSPAC phenotypes we are interested in. We would use this data for exploratory purposes, for selection of appropriate models to be used later for genome-wide SNP association testing. We would also use the data to familiarize ourselves with the ALSPAC resource. The association analysis will be conducted later in collaboration with Drs. David Evans and Nic Timpson. In cases where an association has reached genome wide significance in the 23andMe cohort alone, the ALSPAC data can be used as an independent replication. We also expect that additional loci will be revealed with genome-wide significance in the combined meta-analysis. In some cases, we plan to develop new surveys for the 23andMe cohort around research questions that have been examined in ALSPAC, and will perform meta-analyses for those phenotypes later after new survey data is collected. We think the area of host genetic susceptibility to common infectious disease is particularly interesting and not well studied, despite substantial public health impacts. This includes categorical phenotypes for one-time infections such as measles or mumps, as well as quantitative phenotypes such as frequency of upper respiratory infection and ear ache. We have already identified several associations with infectious disease in the 23andMe customer cohort, and will attempt to replicate these in the ALSPAC cohort. We do not have the resources or domain knowledge to focus a substantial effort on the analysis of any one phenotype. As a result, our plan is to focus on phenotypes that can be assessed from at most a few survey or clinical exam responses, or where summary derived variables are already available. In the phenotype areas for which we have requested data, we have existing survey data and GWAS results in the 23andMe cohort for laterality, bitter taste perception, myopia, some infections, and dental health. We are planning to deploy surveys for post-partum depression, and have requested related ALSPAC data. We have also requested data for coordination and non-verbal perception tests, where we are considering new surveys. Several existing 23andMe surveys measure components of personality, however the metrics we are using are somewhat different from the ones used in ALSPAC. We plan additional work to identify where there is overlap and where we may want to deploy additional surveys in the 23andMe cohort.

There is already a large and growing literature on the effects of month of birth and school starting age on later academic and labour market outcomes. Given the importance of educational attainment in determining a range of later life outcomes - from the probability of being in work and the wage you receive, to health issues and criminal activity - these differences (which arise simply because of the month in which you were born) have the potential to affect individuals throughout their lives. But it is not only educational attainment that has long-lasting effects: there is a body of literature which emphasises the significant effects that a whole range of skills and behaviours developed and exhibited during childhood - including cognitive, non-cognitive, behavioural, social and emotional skills, and anti-social and other risky behaviours (e.g. smoking, drinking and drug use) - may have on later outcomes.

There are a number of reasons why we might expect month of birth (and the age at which you start school and sit academic tests) to affect the development of these skills. For example, whether you enjoy school has been found to be correlated not only with later academic performance, but also with engagement in a range of risky behaviours (including smoking, drinking and cannabis use), all of which may create health costs later in life. Similarly, motivation and perseverance on particular tasks have been found to be significantly positively associated with adult wages, even after taking differences in educational attainment into account. If consistently being amongst the youngest (and perhaps also the smallest) in your class affects your enjoyment of school and/or your motivation and determination to do well (amongst other things), then the month in which you were born - and the age at which you start school and sit academic tests - may have long-term consequences far beyond those captured by educational attainment alone.

Despite these (and other) potentially important repercussions, however, there is currently little evidence available regarding the extent to which month of birth is associated with the development of these types of skills and behaviours. Nor has there been any attempt to investigate whether particular school admissions policies either mitigate or propagate these relationships. The main aims of the proposed programme of research are thus:

1) To identify the impact of month of birth on the development of a range of key skills - including cognitive, non-cognitive, behavioural, social and emotional skills - and engagement in a range of risky behaviours - including smoking, drinking and anti-social behaviour - amongst today's children, from birth through to early adulthood. This work will extend far beyond the scope of previous research in this area - both in terms of the range of skills and behaviours considered, and the ability to consider recent cohorts of children - enabling us to build up a more complete picture of the impact that month of birth has on children's lives within the current policy environment.

2) To identify the best school admissions policy - in terms of all-round skill development and overall behaviour, as well as educational attainment - for a child born towards the end of the academic year, with a view to making clear and practical policy recommendations.

We intend to address our stated research aims using data on three separate contemporary cohorts of children which, together, will enable us to investigate how month of birth and school admissions policies combine to affect skill development and engagement in risky behaviours from birth through to early adulthood. These datasets are the MCS, ALSPAC, and the LSYPE. All three datasets contain excellent information on a range of cognitive and non-cognitive (including behavioural, social and emotional) skills, as well as details of engagement in a variety of risky behaviours. Furthermore, they have also been linked to administrative data on educational attainment.

In addition to data on the range of skills and behaviours that we are interested in, we also need information about the admissions policies that were in place when our cohorts of children started school. This broadly corresponds to 2005-06 or 2006-07 for the MCS, to 1995-96, 1996-97 or 1997-98 for ALSPAC, and to 1994-95 for the LSYPE. In 2004-05, members of the proposed research team started collecting annual admissions policy information from every LEA in England. We also collected retrospective information going back to 1989-90 for the majority of areas. This is a unique data resource, which will enable us to determine the admissions policy regime under which each child started school, and hence to assess which is the "best" admissions policy for children born towards the end of the academic year.

To estimate the impact of month of birth on a wide range of skills and behaviours, we intend to adopt a similar methodology to that used by Crawford, Dearden & Meghir (2007). We describe this methodology here by referring to a comparison of outcomes between August- and September-born children (i.e. those born at the very start and end of the academic year), but it is straightforward to extend this analysis to compare children born in all months at once (which is the approach we would adopt in our research).

In seeking to identify the impact of month of birth on any outcome, one needs to compare individuals who differ only to the extent that they were born in different months. If we are able to assume that date of birth is random, then we may simply compare the outcomes of children born in August with those of children born in September (for example). For date of birth to be random, we need two conditions to hold:

a) Parents with different characteristics - both observed and unobserved (to the researcher) - do not choose to have children in different months.

b) Parents with children born in different months do not differentially respond to survey questionnaires. (This is specific to our reliance on survey data for this project.)

If we are not convinced that these two conditions hold, then it will be necessary to control for all other outcome-relevant characteristics (aside from month of birth) in order to recover the true causal impact of month of birth on particular skills and behaviours. Given the extremely rich nature of the survey data we intend to use for this project (described above), it should be sufficient to do this using a simple linear regression framework.

To go further and identify which admissions policy is "best" for children born at the end of the academic year, we need to make an additional assumption:

c) The admissions policy adopted by the Local Education Authority does not reflect outcome-relevant unobserved differences between areas (or individuals living in areas).

Again, the very wide range of characteristics that we are able to observe in our survey datasets (described above) means that we should be able to simply repeat this analysis for children who start school under different admissions policy regimes, in order to test whether the impact of month of birth differs by admissions policy

Background

Cortical bone density measured using pQCT is a composite of tissue mineral density - which is higher in older, more mature bone - and cortical porosity. Bone resorption may affect both components; if older, denser bone is remodelled and replaced with new bone, overall tissue density will fall. The number of remodelling events, their duration and the size of the intracortical tunnels created will all impact on cortical porosity. Higher levels of bone remodelling in bone diseases such osteogenesis imperfecta are associated with larger cortical pores and degraded bone architecture. Bone remodelling activity is higher during periods of growth; height velocity is strongly associated with fracture risk in childhood, with the peaks of fracture incidence in boys and girls corresponding to the period immediately following maximal height velocity in both sexes. Thus bone resorption and its regulation are key determinants of both short term functional outcomes during skeletal development and of longer term bone mass accrual and bone architecture.

Our previous findings suggest that the RANK/RANKL/OPG system plays an important role in bone development, presumably as a consequence of effects on bone resorption for which this system is a key regulator (Boyle et al). For example, in our previous studies in Sheffield, we found that levels of OPG (a decoy receptor which acts as an endogenous inhibitor of RANKL-dependent osteoclast activation) are reduced in obese children with prior fractures and reduced bone mass (submitted for publication). A trend was also observed towards increased bone mass as OPG increases (t=1.58; unpublished data based on analyses in 103 children, 53 obese). In meta-analyses based on ALSPAC and the GOOD cohort, RANK and OPG polymorphisms previously reported to be associated with BMD in GWAS studies in adults were also associated with cortical BMD at age 16-18 (submitted for publication), as was a newly identified RANKL polymorphism (paper in preparation).

Whereas setting of the RANK/RANKL/OPG system and hence levels of bone resorption are likely to be influenced by constitutional factors such as genetic polymorphisms, this system may also represent the mechanism whereby other factors influence bone resorption. For example, serum OPG is related to fat mass which is in turn related to bone mass, so it may be that external influences to the skeleton like fat mass affect bone mass via perturbations of the RANK/RANKL/OPG system. Furthermore, to the extent that any tendency for fat mass to affect the latter is genetically determined, it may be that RANK/RANKL/OPG polymorphisms contribute to important gene-environmental interactions for skeletal development.

Aims and purpose

We aim to gain an understanding of the role played by bone resorption in skeletal development, and the extent to which this reflects perturbations in the RANK/RANKL/OPG system, by studying the ALSPAC cohort at age 15.5 years in order to address the following questions:-

1. We will investigate whether bone resorption, based on measurement of fasting serum CTx, affects bone mass. CTx levels will be related to total body DXA including the spine sub-region to study trabecular bone, and pQCT of the mid-tibia to examine cortical bone, as measured concurrently.

2. We will establish whether bone resorption is determined by setting of the RANK/RANKL/OPG axis, by examining whether serum CTx is associated with serum RANKL, OPG or their ratio as measured concurrently, and by studying whether RANKL/OPG levels show similar relationships to skeletal development to those observed for CTx levels in (1).

3. External influences on bone resorption and the RANK/RANKL/OPG system will be identified, by studying associations between RANKL/OPG/CTx levels and other factors known to influence skeletal development such as fat and lean mass (DXA at age 15.5), physical activity (accelerometers at age 14), early life factors such as maternal vitamin D status (measured in last trimester in all mothers), childhood vitamin D (measured in all children at 10), and other metabolic factors such as fasting insulin.

4. Constitutional determinants of the RANK/RANKL/OPG system and bone resorption will be determined by identifying genetic markers for RANKL/OPG levels, based on association with RANK/RANKL/OPG polymorphisms as previously analysed in all ALSPAC children in studies to identify genetic determinants of cortical BMD described above, and with genome wide genetic data which is currently available in 3000 ALSPAC children.

5. To establish whether the RANK/RANKL/OPG system contributes to gene-environmental interactions during skeletal development, we will establish whether associations between external factors and RANKL/OPG/CTx levels identified in (3) interact with genetic markers identified in (4).

6. We will investigate whether high levels of CTx or RANKL, or low levels of OPG, or genetic determinants thereof, can be used to identify children at increased risk of fracture, either alone or in combination.

Summary of plan of investigation

This project will be based on the 2940 children who attended the ALSPAC research clinic at age 15.5 years and had fasting blood samples collected and underwent total body DXA scans. The present proposal is intended to cover the cost of transferring an aliquot of these samples to Sheffield and subsequent analysis of serum RANKL, OPG and CTx by validated immunoassay. Error checking and descriptive statistics will be performed by a 50% time research assistant based at ALSPAC funded by this project. After linkage to the ALSPAC database, these serum levels will be analysed in relation to results collected at the same research clinic ie pQCT scans, total body DXA scans, fasting insulin. These data will also be analysed in relation to RANK/RANKL/OPG genetic polymorphisms as previously analysed in all available children, and to available child GWAS data. RANKL, OPG and CTx levels will also be analysed for associations with other potential influences on skeletal metabolism based on measures collected at other time points, such as habitual level of physical data (Actigraph data at age 14) and vitamin D levels (third trimester mothers, children aged 10).

Background and objective - Several studies have reported an increased risk of breast cancer in twins compared with singletons (1-3). The risk appears to be larger in dizygotic compared to monozygotic twins and singletons. Dizygotic pregnancies may have two placentas or one larger placenta which may present opportunities for larger estrogen exposures than monozygotic pregnancies (4). Early age at menarche is also linked to an increase in the risk of breast cancer later in life (5). We would like to explore whether twinship, zygosity, and the sex of the co-twin are associated with timing of puberty in girls enrolled in the Avon Longitudinal Study of Parents and Children.

Statistical analyses - Bivariate analyses will be conducted to test maternal prenatal characteristics. These will be retained in the analyses as confounders if they are associated to the study outcomes. We will construct parametric survival models to estimate the adjusted difference in median age of entry into stages >=2 of breast and pubic hair development and menarche associated with the predictors under examination in comparison groups. Analyses will be conducted using PROC LIFEREG (6) in SAS. Human subject protection was assessed and approved by the ALSPAC Law and Ethics Committee, the Local Research Ethics Committees, and the CDC Institutional.

Preliminary analyses - Of the ~ 4000 girls enrolled in ALSPAC who responded to the Growing and Changing Questionnaire I identified 82 co-twin girls with valid puberty data. Of these: 1) 44 girls (22 pairs) have a female co-twin with valid puberty data; 2) another 6 girls have a female co-twin with invalid puberty data; and 3) 31 have girls a male co-twin. The number of co-twin girls with puberty data appears to be adequate to conduct the analyses proposed.

Data needs from ALSPAC - 1) birth weight of twin births; and 2) zygosity of the twins.

Summary

Overactive bladder (OAB) affects approximately 13% of adults [1], with a major impact on quality of life, equivalent to diabetes or rheumatoid arthritis [2, 3]. There is an urgent need for objective biomarkers to help us predict occurrence, progression, and therapeutic response. beta3-adrenoceptor gene (ADRb3), transient receptor potential vanilloid 4 (TRPV4) and nerve growth factor beta polypeptide gene (NGFB) single nucleotide polymorphisms (SNPs) together represent highly plausible candidates as genetic population risk factors for OAB. The initial phase of this study will use existing DNA, banked from the ALSPAC cohort. Using mailings of a contemporary validated self-completion questionnaire, to women in the sample, the current prevalence of OAB will be established. Genotyping will be performed for all participants (including men) for 41 SNPs (see appendix), characterising all known common variation around the ADRb3, TRPV4 and NGFB loci. The study will assess the relationship of these variants to OAB prevalence, OAB onset, and OAB severity, while controlling for known risk factors for OAB. The large size of the biobank, and the detailed follow-up of the sample over the last 20 years, provides a unique opportunity to assess the relationship of genetics and the environment in the natural history of OAB. The data should allow us to predict disease susceptibility, lead to improved treatment with current agents, and provide evidence for new potential drug targets. We therefore believe the study will have dramatic impact on our ability to diagnose and treat patients suffering with OAB.

Background

Cross-sectional studies [4] and classical twin studies [5] have suggested that OAB is highly heritable, with an affected first degree relative being associated with 1.3-1.9 fold increased risk, however no candidate genes have been identified. This study proposes to investigate three potential genetic biomarkers. beta3-adrenoceptors are the principle mediator of detrusor relaxation [6], and the YM-178 beta3-adrenoceptor agonist is in development for treatment of OAB [7]. The beta3-adrenoceptor gene (ADRb3) has several common SNPs, including the rs4994 polymorphism, which is associated with diabetes and obesity. In a single small sample is has also reported to be more common in women with OAB (OR 3.0, RR 2.1, n=201) [8], indicating the possibility of a shared susceptibility to OAB and metabolic syndrome. TRPV4 functions as a mechano-sensitive channel in the urothelium, with TRPV4 agonists inducing bladder overactivity in mice [9]. It is considered a key target for future OAB drug development. Functional TRPV4 SNPs have diverse pathological associations, including obstructive airways disease [10], and hyponatraemia [11]. NGF is one link between tissue inflammation and bladder C-fiber afferent nerve excitability. Urinary NGF levels correlate with multiple measures of OAB disease severity [12], and demonstrate responsiveness to therapy. A common functional SNP of the NGFB gene, rs6330, is associated with conditions linked to OAB, including anxiety traits [13] and multiple sclerosis [14]. ADRb3, TRPV4 and NGFB SNPs therefore together represent highly plausible candidates as genetic population risk factors, and require investigation in a population based sample.

Objectives

This study will characterise all known common variation around the ADRb3, TRPV4, and NGFB loci, with particular interest in the rs4994, rs12578401 and rs6330 variants, in a population based study of OAB symptoms in women. We will identify the role of ADRb3, TRPV4, and NGF SNPs and haplotypes in the pathophysiology of OAB.

Main hypothesis:

Common functional variation of the ADRb3, TRPV4 and NGFB genes is associated with increased risk of OAB in women.

Secondary hypotheses:

Common functional variation of the ADRb3, TRPV4 and NGFB genes is associated with alterations in disease phenotype in women with OAB:

* earlier onset of adult symptoms,

* later remission of childhood enuresis

* increased subjective severity and bother of OAB symptoms

Common functional variation of the ADRb3, TRPV4 and NGFB genes modifies environmental risks for OAB, including parity, smoking, weight gain and dietary habits.

Methodology

Design: Population based genetic association study of approximately 15,000 mothers and daughters.

Genotyping: KBiosciences will genotype 41 SNPs (see appendix) around the ADRb3, TRPV4 and NGFB genes.

Case assessment: Up to 3 mailings of the 12 item validated International Consultation on Incontinence - Female Lower Urinary Tract Symptoms Questionnaire, for self completion, with OAB cases identified using item 5a: greater than = "nine to ten" voids per day, and item 3a: "sometimes" "most of the time" or "all of the time" sudden need to rush to the toilet to urinate. Existing items in the ALSPAC database on urinary symptoms in mothers, asking about incontinence and frequency will be used to assess the onset of OAB, and items about bed wetting in children will be used to assess the remission of childhood enuresis.

Assessment of confounders: The existing database includes 1-2 yearly details from 1991 for mothers and daughters of birth records, co-morbidities, medication use, dietary habits (food, alcohol, caffeine), smoking, socioeconomics, ethnicity, and weight gain.

Statistics: Genotype data will be analysed using Hardy-Weinberg Equilibrium testing, and Cochran Armitage trend test with input and advice from statistical geneticists. The sample size is powered on the basis of the minor allele frequencies for the rs4994 SNP. With the frequency estimated at 10%, and alpha set at 0.05, an estimated sample size of 10000 provides 90% power to detect a 3% difference in the allele prevalence.

(No proposal form received).

The nature versus nurture distinction, while traditional, is obsolete. The modern literature on epigenetic expression and gene-environment interactions shows that the sharp distinction between acquired skills and ability featured in the early human capital literature and that still guides much research in health economics (Grossman 1972, 2000) is not tenable (Pray 2004, Rutter 2006). Additive "nature" and "nurture" models, while traditional and still used in many studies of heritability and family influence, mischaracterize how ability and health are manifested. Genes and environment cannot be meaningfully distinguished by traditional linear models that assign unique variances to each component. Abilities are produced, and gene expression is governed, by environmental conditions (see Rutter 2006). Measured abilities are the outcome of environmental influences, including in utero experiences, and also have genetic components.

Behaviors and abilities have both a genetic and an acquired character. Recent analyses in economics that break the "causes" of birthweight into environmental and genetic components ignore the lessons of the recent epigenetic and gene-environment literatures, and provide misleading guides to policy because they suggest the operation of "genetic" factors are beyond the scope of policy intervention. We seek to expand our studies beyond the United States by using the data to address the following:

(1) What environmental (parental, neighborhood) inputs affect health, cognition and socioemotional skills at different stages of childhood? What are the critical and sensitive periods for remediation?

(2) How do these stocks of capacities cross-fertilize each other? How effective are socioemotional and cognitive interventions on health and its evolution? How do health stocks cross-fertilize socioemotional and cognitive skill production?

(3) What are the costs and the benefits of interventions for remediation and prevention at different stages of the lifecycle of children?

Background:

In observational studies on causality, including non-randomised therapeutic studies, the exposure and confounders are often considered constant over time. However, this assumption may not be valid, especially in case of time-varying confounding [Robins 1992, Robins 2000]. When exposure status affects confounder status, and this confounder again affects future exposure, conditioning on the confounder (as is done in a regular regression analysis), might induce bias rather than correct for confounding. To adequately estimate causal effects of e.g. interventions, initiation or stopping of medical therapy during follow-up should be taken into account, for example using time-dependent analyses. Several time-dependent analyses have been proposed, such as Cox proportional hazards analysis with time-dependent exposure, GEE, and inverse probability of treatment weighting (IPTW). In empirical data (observational or trials), however, these have not been compared throughout. Furthermore, theoretically the method could give different results, for example in case of time/varying confounding. However, to what extent time/varying confounding will affect studies over a wide range of clinical fields, is unclear. We therefore, aim to compare analyses for time-varying exposure for the association between breastfeeding and wheeze in two examples on observational studies (the Whistler cohort and the ALSPAC cohort), and compare the results to those from a randomised trial.

Objective:

To compare methods to assess time-varying exposure in observational studies on the effects of breastfeeding on asthma/eczema, and compare the associations found to those from a randomised controlled trial.

Research questions:

1. What is the association between breastfeeding and wheezing? (potential confounders: maternal smoking, being in a smoky room, daycare, maternal wheeze/eczema, age, weight and height, pets)

2. What is the association between breastfeeding and eczema? (potential confounders: maternal smoking, being in a smoky room, daycare, maternal wheeze/eczema, age, weight and height, pets)

Methods:

Several time-varying analyses will be applied and compared:

1) Exposure = ever vs. never breastfed (dichotomous measure);

2) Exposure = total duration of breastfeeding (i.e.,exposure is a continuous measure);

3) Exposure is time-dependent (i.e. breastfeeding at different ages of follow-up) and effects will be estimated using survival analysis with this time-dependent exposure. Confounders are considered constant over time (e.g. maternal smoking during pregnancy is taken as the constant value for the confounder "maternal smoking");

4) Exposure and confounders are both time-dependent. Effects are estimated using survival analysis.Follow-up is divided in discrete intervals (i.e., 1-year intervals). Each period is considered an (independent) observation. Covariates can differ at the baseline of each period.

5) Exposure and confounders are both time-dependent. Effects are estimated using Generalized Estimating Equations, thus taking dependency between observations into account.

6) Exposure and confounders are both time-dependent. Effects are estimated using marginal structural models using inverse probability of treatment weighting [Robins 2000]

7) Exposure and confounders are both time-dependent. Effects are estimated using G-estimation [Tilling 2002].

We propose to evaluate data being gathered at the 17 Clinic using existing forms, most of which were developed through a collaboration between our group at UCSD and investigators at ALSPAC. Our goal is to use our expertise with SEM and with LR-based mediational model to analyse existing ALSPAC data to enhance our understanding of how mediators of LR to outcome function in the mid-teens. Eventually, we hope these analyses will contribute to our ability to develop programs to help diminish the risk for heavy drinking and alcohol problems in teenagers carrying a risk for these adverse alcohol outomes through a low LR.