Ignoring temporal trends in genetic effects substantially reduces power of quantitative trait linkage analysis

Genetic and Environmental Influences on the Developmental Course of Attention-Deficit/Hyperactivity Disorder Symptoms From Childhood to Adolescence

IMPORTANCE

Attention-deficit/hyperactivity disorder (ADHD) is conceptualized as a neurodevelopmental disorder that is strongly heritable. However, to our knowledge, no study to date has examined the genetic and environmental influences explaining interindividual differences in the developmental course of ADHD symptoms from childhood to adolescence (ie, systematic decreases or increases with age). The reason ADHD symptoms persist in some children but decline in others is an important concern, with implications for prognosis and interventions.

OBJECTIVE

To assess the proportional impact of genes and the environment on interindividual differences in the developmental course of ADHD symptom domains of hyperactivity/impulsivity and inattention between ages 8 and 16 years.

DESIGN, SETTING, AND PARTICIPANTS

A prospective sample of 8395 twin pairs from the Twins Early Development Study, recruited from population records of births in England and Wales between January 1, 1994, and December 31, 1996. Data collection at age 8 years took place between November 2002 and November 2004; data collection at age 16 years took place between February 2011 and January 2013.

MAIN OUTCOMES AND MEASURES

Both DSM-IV ADHD symptom subscales were rated 4 times by participants' mothers.

RESULTS

Estimates from latent growth curve models indicated that the developmental course of hyperactivity/impulsivity symptoms followed a sharp linear decrease (mean score of 6.0 at age 8 years to 2.9 at age 16 years). Interindividual differences in the linear change in hyperactivity/impulsivity were under strong additive genetic influences (81%; 95% CI, 73%-88%). More than half of the genetic variation was specific to the developmental course and not shared with the baseline level of hyperactivity/impulsivity. The linear decrease in inattention symptoms was less pronounced (mean score of 5.8 at age 8 years to 4.9 at age 16 years). Nonadditive genetic influences accounted for a substantial amount of variation in the developmental course of inattention symptoms (54%; 95% CI, 8%-76%), with more than half being specific to the developmental course.

CONCLUSIONS AND RELEVANCE

The large genetic influences on the developmental course of ADHD symptoms are mostly specific and independent of those that account for variation in the baseline level of symptoms. Different sets of genes may be associated with the developmental course vs the baseline level of ADHD symptoms and explain why some children remit from ADHD, whereas others persist. Recent longitudinal imaging data indicate that the maintenance or increase in symptoms is underpinned by atypical trajectories of cortical development. This may reflect a specific genetic liability, distinct from that which contributes to baseline ADHD symptoms, and warrants closer follow-up.

Linkage analysis incorporating gene-age interactions identifies seven novel lipid loci: the Family Blood Pressure Program

OBJECTIVE

To detect novel loci with age-dependent effects on fasting (≥ 8 h) levels of total cholesterol, high-density lipoprotein, low-density lipoprotein, and triglycerides using 3600 African Americans, 1283 Asians, 3218 European Americans, and 2026 Mexican Americans from the Family Blood Pressure Program (FBPP).

METHODS

Within each subgroup (defined by network, race, and sex), we employed stepwise linear regression (retention p ≤ 0.05) to adjust lipid levels for age, age-squared, age-cubed, body-mass-index, current smoking status, current drinking status, field center, estrogen therapy (females only), as well as antidiabetic, antihypertensive, and antilipidemic medication use. For each trait, we pooled the standardized male and female residuals within each network and race and fit a generalized variance components model that incorporated gene-age interactions. We conducted FBPP-wide and race-specific meta-analyses by combining the p-values of each linkage marker across subgroups using a modified Fisher's method.

RESULTS

We identified seven novel loci with age-dependent effects; four total cholesterol loci from the meta-analysis of Mexican Americans (on chromosomes 2q24.1, 4q21.21, 8q22.2, and 12p11.23) and three high-density lipoprotein loci from the meta-analysis of all FBPP subgroups (on chromosomes 1p12, 14q11.2, and 21q21.1). These loci lacked significant genome-wide linkage or association evidence in the literature and had logarithm of odds (LOD) score ≥ 3 in the meta-analysis with LOD ≥ 1 in at least two network and race subgroups (exclusively of non-European descent).

CONCLUSION

Incorporating gene-age interactions into the analysis of lipids using multi-ethnic cohorts can enhance gene discovery. These interaction loci can guide the selection of families for sequencing studies of lipid-associated variants.

Hierarchical linear modeling of longitudinal pedigree data for genetic association analysis

Genetic association analysis on complex phenotypes under a longitudinal design involving pedigrees encounters the problem of correlation within pedigrees, which could affect statistical assessment of the genetic effects. Approaches have been proposed to integrate kinship correlation into the mixed-effect models to explicitly model the genetic relationship. These have proved to be an efficient way of dealing with sample clustering in pedigree data. Although current algorithms implemented in popular statistical packages are useful for adjusting relatedness in the mixed modeling of genetic effects on the mean level of a phenotype, they are not sufficiently straightforward to handle the kinship correlation on the time-dependent trajectories of a phenotype. We introduce a 2-level hierarchical linear model to separately assess the genetic associations with the mean level and the rate of change of a phenotype, integrating kinship correlation in the analysis. We apply our method to the Genetic Analysis Workshop 18 genome-wide association studies data on chromosome 3 to estimate the genetic effects on systolic blood pressure measured over time in large pedigrees. Our method identifies genetic variants associated with blood pressure with estimated inflation factors of 0.99, suggesting that our modeling of random effects efficiently handles the genetic relatedness in pedigrees. Application to simulated data captures important variants specified in the simulation. Our results show that the method is useful for genetic association studies in related samples using longitudinal design.

Bivariate association analysis of longitudinal phenotypes in families

Statistical genetic methods incorporating temporal variation allow for greater understanding of genetic architecture and consistency of biological variation influencing development of complex diseases. This study proposes a bivariate association method jointly testing association of two quantitative phenotypic measures from different time points. Measured genotype association was analyzed for single-nucleotide polymorphisms (SNPs) for systolic blood pressure (SBP) from the first and third visits using 200 simulated Genetic Analysis Workshop 18 (GAW18) replicates. Bivariate association, in which the effect of an SNP on the mean trait values of the two phenotypes is constrained to be equal for both measures and is included as a covariate in the analysis, was compared with a bivariate analysis in which the effect of an SNP was estimated separately for the two measures and univariate association analyses in 9 SNPs that explained greater than 0.001% SBP variance over all 200 GAW18 replicates.The SNP 3_48040283 was significantly associated with SBP in all 200 replicates with the constrained bivariate method providing increased signal over the unconstrained bivariate method. This method improved signal in all 9 SNPs with simulated effects on SBP for nominal significance (p-value <0.05). However, this appears to be determined by the effect size of the SNP on the phenotype. This bivariate association method applied to longitudinal data improves genetic signal for quantitative traits when the effect size of the variant is moderate to large.

Body mass index modulates blood pressure heritability: the Family Blood Pressure Program

BACKGROUND

Candidate gene and twin studies suggest that interactions between body mass index (BMI) and genes contribute to the variability of blood pressure (BP). To determine whether there is evidence for gene-BMI interactions, we investigated the modulation of BP heritability by BMI using 4,153 blacks, 1,538 Asians, 4,013 whites, and 2,199 Hispanic Americans from the Family Blood Pressure Program.

METHODS

To capture the BP heritability dependence on BMI, we employed a generalized variance components model incorporating linear and Gaussian interactions between BMI and the genetic component. Within each race and network subgroup, we used the Akaike information criterion and likelihood ratio test to select the appropriate interaction function for each BP trait (systolic BP (SBP), diastolic BP (DBP), mean arterial pressure (MAP), and pulse pressure (PP)) and determine interaction significance, respectively.

RESULTS

BP heritabilities were significantly modified by BMI in the GenNet and SAPPHIRe Networks, which contained the youngest and least-obese participants, respectively. GenNet Whites had unimodal SBP, MAP, and PP heritabilities that peaked between BMI values of 33 and 37kg/m(2). The SBP and MAP heritabilities in GenNet Hispanic Americans, as well as the PP heritability in GenNet blacks, were increasing functions of BMI. The DBP and SBP heritabilities in the SAPPHIRe Chinese and Japanese, respectively, were decreasing functions of BMI.

CONCLUSIONS

BP heritability differed by BMI in the youngest and least-obese networks, although the shape of this dependence differed by race. Use of nonlinear gene-BMI interactions may enhance BP gene discovery efforts in individuals of European ancestry.

Longitudinal association analysis of quantitative traits

Longitudinal genetic studies provide a valuable resource for exploring key genetic and environmental factors that affect complex traits over time. Genetic analysis of longitudinal data that incorporate temporal variations is important for understanding genetic architecture and biological variations of common complex diseases. Although they are important, there is a paucity of statistical methods to analyze longitudinal human genetic data. In this article, longitudinal methods are developed for temporal association mapping to analyze population longitudinal data. Both parametric and nonparametric models are proposed. The models can be applied to multiple diallelic genetic markers such as single-nucleotide polymorphisms and multiallelic markers such as microsatellites. By analytical formulae, we show that the models take both the linkage disequilibrium and temporal trends into account simultaneously. Variance-covariance structure is constructed to model the single measurement variation and multiple measurement correlations of an individual based on the theory of stochastic processes. Novel penalized spline models are used to estimate the time-dependent mean functions and regression coefficients. The methods were applied to analyze Framingham Heart Study data of Genetic Analysis Workshop (GAW) 13 and GAW 16. The temporal trends and genetic effects of the systolic blood pressure are successfully detected by the proposed approaches. Simulation studies were performed to find out that the nonparametric penalized linear model is the best choice in fitting real data. The research sheds light on the important area of longitudinal genetic analysis, and it provides a basis for future methodological investigations and practical applications.

Flexible semiparametric analysis of longitudinal genetic studies by reduced rank smoothing

In family-based longitudinal genetic studies, investigators collect repeated measurements on a trait that changes with time along with genetic markers. Since repeated measurements are nested within subjects and subjects are nested within families, both the subject-level and measurement-level correlations must be taken into account in the statistical analysis to achieve more accurate estimation. In such studies, the primary interests include to test for quantitative trait locus (QTL) effect, and to estimate age-specific QTL effect and residual polygenic heritability function. We propose flexible semiparametric models along with their statistical estimation and hypothesis testing procedures for longitudinal genetic designs. We employ penalized splines to estimate nonparametric functions in the models. We find that misspecifying the baseline function or the genetic effect function in a parametric analysis may lead to substantially inflated or highly conservative type I error rate on testing and large mean squared error on estimation. We apply the proposed approaches to examine age-specific effects of genetic variants reported in a recent genome-wide association study of blood pressure collected in the Framingham Heart Study.

Enriching rare variants using family-specific linkage information

Genome-wide association studies have been successful in identifying common variants for common complex traits in recent years. However, common variants have generally failed to explain substantial proportions of the trait heritabilities. Rare variants, structural variations, and gene-gene and gene-environment interactions, among others, have been suggested as potential sources of the so-called missing heritability. With the advent of exome-wide and whole-genome next-generation sequencing technologies, finding rare variants in functionally important sites (e.g., protein-coding regions) becomes feasible. We investigate the role of linkage information to select families enriched for rare variants using the simulated Genetic Analysis Workshop 17 data. In each replicate of simulated phenotypes Q1 and Q2 on 697 subjects in 8 extended pedigrees, we select one pedigree with the largest family-specific LOD score. Across all 200 replications, we compare the probability that rare causal alleles will be carried in the selected pedigree versus a randomly chosen pedigree. One example of successful enrichment was exhibited for gene VEGFC. The causal variant had minor allele frequency of 0.0717% in the simulated unrelated individuals and explained about 0.1% of the phenotypic variance. However, it explained 7.9% of the phenotypic variance in the eight simulated pedigrees and 23.8% in the family that carried the minor allele. The carrier’s family was selected in all 200 replications. Thus our results show that family-specific linkage information is useful for selecting families for sequencing, thus ensuring that rare functional variants are segregating in the sequencing samples.

Semiparametric variance components models for genetic studies with longitudinal phenotypes

In a family-based genetic study such as the Framingham Heart Study (FHS), longitudinal trait measurements are recorded on subjects collected from families. Observations on subjects from the same family are correlated due to shared genetic composition or environmental factors such as diet. The data have a 3-level structure with measurements nested in subjects and subjects nested in families. We propose a semiparametric variance components model to describe phenotype observed at a time point as the sum of a nonparametric population mean function, a nonparametric random quantitative trait locus (QTL) effect, a shared environmental effect, a residual random polygenic effect and measurement error. One feature of the model is that we do not assume a parametric functional form of the age-dependent QTL effect, and we use penalized spline-based method to fit the model. We obtain nonparametric estimation of the QTL heritability defined as the ratio of the QTL variance to the total phenotypic variance. We use simulation studies to investigate performance of the proposed methods and apply these methods to the FHS systolic blood pressure data to estimate age-specific QTL effect at 62cM on chromosome 17.

Flexible estimation of covariance function by penalized spline with application to longitudinal family data

Longitudinal data are routinely collected in biomedical research studies. A natural model describing longitudinal data decomposes an individual's outcome as the sum of a population mean function and random subject-specific deviations. When parametric assumptions are too restrictive, methods modeling the population mean function and the random subject-specific functions nonparametrically are in demand. In some applications, it is desirable to estimate a covariance function of random subject-specific deviations. In this work, flexible yet computationally efficient methods are developed for a general class of semiparametric mixed effects models, where the functional forms of the population mean and the subject-specific curves are unspecified. We estimate nonparametric components of the model by penalized spline (P-spline, Biometrics 2001; 57:253-259), and reparameterize the random curve covariance function by a modified Cholesky decomposition (Biometrics 2002; 58:121-128) which allows for unconstrained estimation of a positive-semidefinite matrix. To provide smooth estimates, we penalize roughness of fitted curves and derive closed-form solutions in the maximization step of an EM algorithm. In addition, we present models and methods for longitudinal family data where subjects in a family are correlated and we decompose the covariance function into a subject-level source and observation-level source. We apply these methods to the multi-level Framingham Heart Study data to estimate age-specific heritability of systolic blood pressure nonparametrically.

Curve-based multivariate distance matrix regression analysis: application to genetic association analyses involving repeated measures

Most, if not all, human phenotypes exhibit a temporal, dosage-dependent, or age effect. Despite this fact, it is rare that data are collected over time or in sequence in relevant studies of the determinants of these phenotypes. The costs and organizational sophistication necessary to collect repeated measurements or longitudinal data for a given phenotype are clearly impediments to this, but greater efforts in this area are needed if insights into human phenotypic expression are to be obtained. Appropriate data analysis methods for genetic association studies involving repeated or longitudinal measures are also needed. We consider the use of longitudinal profiles obtained from fitted functions on repeated data collections from a set of individuals whose similarities are contrasted between sets of individuals with different genotypes to test hypotheses about genetic influences on time-dependent phenotype expression. The proposed approach can accommodate uncertainty of the fitted functions, as well as weighting factors across the time points, and is easily extended to a wide variety of complex analysis settings. We showcase the proposed approach with data from a clinical study investigating human blood vessel response to tyramine. We also compare the proposed approach with standard analytic procedures and investigate its robustness and power via simulation studies. The proposed approach is found to be quite flexible and performs either as well or better than traditional statistical methods.

The impact of newly identified loci on coronary heart disease, stroke and total mortality in the MORGAM prospective cohorts

Recently, genome wide association studies (GWAS) have identified a number of single nucleotide polymorphisms (SNPs) as being associated with coronary heart disease (CHD). We estimated the effect of these SNPs on incident CHD, stroke and total mortality in the prospective cohorts of the MORGAM Project. We studied cohorts from Finland, Sweden, France and Northern Ireland (total N=33,282, including 1,436 incident CHD events and 571 incident stroke events). The lead SNPs at seven loci identified thus far and additional SNPs (in total 42) were genotyped using a case-cohort design. We estimated the effect of the SNPs on disease history at baseline, disease events during follow-up and classic risk factors. Multiple testing was taken into account using false discovery rate (FDR) analysis. SNP rs1333049 on chromosome 9p21.3 was associated with both CHD and stroke (HR=1.20, 95% CI 1.08-1.34 for incident CHD events and 1.15, 0.99-1.34 for incident stroke). SNP rs11670734 (19q12) was associated with total mortality and stroke. SNP rs2146807 (10q11.21) showed some association with the fatality of acute coronary event. SNP rs2943634 (2q36.3) was associated with high density lipoprotein (HDL) cholesterol and SNPs rs599839, rs4970834 (1p13.3) and rs17228212 (15q22.23) were associated with non-HDL cholesterol. SNPs rs2943634 (2q36.3) and rs12525353 (6q25.1) were associated with blood pressure. These findings underline the need for replication studies in prospective settings and confirm the candidacy of several SNPs that may play a role in the etiology of cardiovascular disease.

Genetic effect on blood pressure is modulated by age: the Hypertension Genetic Epidemiology Network Study

Genome-wide linkage analysis was performed for systolic and diastolic blood pressures in the Hypertension Genetic Epidemiology Network. We investigated the role of gene-age interactions using a recently developed variance components method that incorporates age variation in genetic effects. Substantially improved linkage evidence, in terms of both the number of linkage peaks and their significance levels, was observed. Twenty-six linkage peaks were identified with maximum logarithm of odds scores ranging between 3.0 and 4.6, 15 of which were cross-validated by the literature. The chromosomal region 1p36 that showed the highest logarithm of odds score in our study was found to be supported by evidence from 3 studies. The new method also led to vastly improved validation across ethnic groups. Ten of the 15 supported linkage peaks were cross-validated between 2 different ethnic groups, and 2 peaks on chromosomal region 1q31 and 16p11 were validated in 3 ethnic groups. In conclusion, this investigation demonstrates that genetic effects on blood pressure vary by age. The improved genetic linkage results presented here should help to identify the specific genetic variants that explain the observed results.