Estimation of pleiotropy between complex diseases using single-nucleotide polymorphism-derived genomic relationships and restricted maximum likelihood

The University of Queensland, Queensland Brain Institute, Brisbane, QLD 4072, The University of Queensland Diamantina Institute, Princess Alexandra Hospital, Brisbane, QLD 4102 and Department of Agriculture and Food Systems, University of Melbourne, VIC 3010, Melbourne, Australia.
Bioinformatics (Impact Factor: 4.62). 07/2012; 28(19):2540-2. DOI: 10.1093/bioinformatics/bts474
Source: PubMed

ABSTRACT Genetic correlations are the genome-wide aggregate effects of causal variants affecting multiple traits. Traditionally, genetic correlations between complex traits are estimated from pedigree studies, but such estimates can be confounded by shared environmental factors. Moreover, for diseases, low prevalence rates imply that even if the true genetic correlation between disorders was high, co-aggregation of disorders in families might not occur or could not be distinguished from chance. We have developed and implemented statistical methods based on linear mixed models to obtain unbiased estimates of the genetic correlation between pairs of quantitative traits or pairs of binary traits of complex diseases using population-based case-control studies with genome-wide single-nucleotide polymorphism data. The method is validated in a simulation study and applied to estimate genetic correlation between various diseases from Wellcome Trust Case Control Consortium data in a series of bivariate analyses. We estimate a significant positive genetic correlation between risk of Type 2 diabetes and hypertension of ~0.31 (SE 0.14, P = 0.024).
Our methods, appropriate for both quantitative and binary traits, are implemented in the freely available software GCTA ( Supplementary Information: Supplementary data are available at Bioinformatics online.

1 Follower
  • Source
    • "Some research (Conley et al., 2014) suggests that GCTA results may be robust to violations of this assumption, but more work in this area is needed. This univariate GCTA model has been recently extended to a bivariate specification in which the same genetic relationship can be used to examine covariance between two traits as a function of genetic similarity (Lee et al., 2012). The model is derived from twin-sibling approaches in which cross-twin-cross-trait covariance is compared among monozygotic and dizygotic twin pairs to see if there is evidence that the covariance is due to common genetic influences. "
    [Show abstract] [Hide abstract]
    ABSTRACT: We use genome wide data from respondents of the Health and Retirement Study (HRS) to evaluate the possibility that common genetic influences are associated with education and three health outcomes: depression, self-rated health, and body mass index. We use a total of 1.7 million single nucleotide polymorphisms obtained from the Illumina HumanOmni2.5-4v1 chip from 4233 non-Hispanic white respondents to characterize genetic similarities among unrelated persons in the HRS. We then used the Genome Wide Complex Trait Analysis (GCTA) toolkit, to estimate univariate and bivariate heritability. We provide evidence that education (h(2) = 0.33), BMI (h(2) = 0.43), depression (h(2) = 0.19), and self-rated health (h(2) = 0.18) are all moderately heritable phenotypes. We also provide evidence that some of the correlation between depression and education as well as self-rated health and education is due to common genetic factors associated with one or both traits. We find no evidence that the correlation between education and BMI is influenced by common genetic factors.
    Social Science & Medicine 08/2014; 127. DOI:10.1016/j.socscimed.2014.08.001 · 2.56 Impact Factor
  • Source
    • "The univariate GCTA estimates for g have been reported previously (Marioni et al., in press). Bivariate GCTA models (Lee et al. 2012) were then run to obtain estimates of the genetic correlation and bivariate heritability between height and g. "
    [Show abstract] [Hide abstract]
    ABSTRACT: Greater height and higher intelligence test scores are predictors of better health outcomes. Here, we used molecular (single-nucleotide polymorphism) data to estimate the genetic correlation between height and general intelligence (g) in 6,815 unrelated subjects (median age 57, IQR 49–63) from the Generation Scotland: Scottish Family Health Study cohort. The phenotypic correlation between height and g was 0.16 (SE 0.01). The genetic correlation between height and g was 0.28 (SE 0.09) with a bivariate heritability estimate of 0.71. Understanding the molecular basis of the correlation between height and intelligence may help explain any shared role in determining health outcomes. This study identified a modest genetic correlation between height and intelligence with the majority of the phenotypic correlation being explained by shared genetic influences. Electronic supplementary material The online version of this article (doi:10.1007/s10519-014-9644-z) contains supplementary material, which is available to authorized users.
    Behavior Genetics 02/2014; 44(2). DOI:10.1007/s10519-014-9644-z · 2.84 Impact Factor
  • Source
    • "On the other hand, the NR algorithm is fast to converge given a good initial value (only takes a few iterations), but can easily fail to do so given a bad starting point. Therefore, we follow previous approaches [31] [32] [30] [29] [4] and combine the two algorithms together, with the PX-EM algorithm providing a good starting value for the following NR algorithm. In addition, for moderate d the PX-EM algorithm is considerably faster than the NR algorithm, and so for GWAS applications, we perform the NR algorithm only for markers where the p value after the PX-EM algorithm is < 1.0 × 10 −3 (or a user adjusted threshold). "
    [Show abstract] [Hide abstract]
    ABSTRACT: Multivariate linear mixed models (mvLMMs) are powerful tools for testing associations between single-nucleotide polymorphisms and multiple correlated phenotypes while controlling for population stratification in genome-wide association studies. We present efficient algorithms in the genome-wide efficient mixed model association (GEMMA) software for fitting mvLMMs and computing likelihood ratio tests. These algorithms offer improved computation speed, power and P-value calibration over existing methods, and can deal with more than two phenotypes.
    Nature Methods 02/2014; 11(4). DOI:10.1038/nmeth.2848 · 25.95 Impact Factor
Show more