Table 1.
Comparisons of sample sizes, number of non-zero off-diagonal entries and number of pairwise relationships of different degrees between GS10K and GS20K.
Fig 1.
Illustration of the model and matrices.
The diagram shows the relationship between the tested genetic/environmental effect and the individuals in an example pedigree. Each colour represents a specific effect and individuals affected by that effect are circled with that colour. People in grey or black are the people not in or in the data. Examples of how the relationship matrices for those effects look are also given.
Fig 2.
Boxplots for estimates of each component obtained from models ‘G’, ‘K’, ‘F’, ‘S’, ‘C’, ‘GK’, ‘GKC’, ‘GKSC’ and ‘GKFSC’.
X-axis: the contributors to the simulated phenotype and the model used (matched model); Y-axis: proportion of total phenotypic variance captured by each design matrix. Yellow lines: simulated value for each component. Parameter settings: = 0.3,
= 0.2,
= 0.1,
= 0.1 and
= 0.05. For example, the 2nd boxplot of the 3rd graph means that, the simulated phenotypes are contributed by 30%, 20%, 10% and 40% of SNP-associated, pedigree-associated, couple environmental and residual effects respectively; we conducted variance component analyses for all replicates using the matched model ‘GKC’ and the estimates of
range from about 8% to 12% with a mean of 10%, as expected.
Fig 3.
Heritability estimates using subpopulations of GS10K with different GRM cut-off points.
X-axis: the number of individuals among whom the pairwise relationship is larger than a specified degree; Y-axis: Heritability estimates with ± 2 s.e.
Table 2.
Results of variance component analyses for anthropometric and cardiometabolic traits using final models selected from the stepwise model selection and the full model in GS10K.
Table 3.
Results of variance component analyses for anthropometric and cardiometabolic traits using final models selected from the stepwise model selection and the full model in GS20K.
Fig 4.
Results of variance component analysis using final selected models for anthropometric and cardiometabolic traits in GS20K.
X-axis: names of phenotype; Y-axis: proportion of phenotypic/genetic variance explained by the different components. a) Proportion of phenotypic variance explained by genetics and environment for each trait. b) Proportion of phenotypic variance explained by different components kept in the selected model for each trait. c) Proportion of genetic variance explained by SNP-associated and pedigree-associated genetic effects.
Table 4.
Comparisons of the results from final models in GS20K to previous published results.