Figure 1.
Trajectories of phenotype left hippocampus volume over time (in months) in three allele groups of SNP rs2075650.
Table 1.
Significant SNPs and each one's associated phenotype numbers at the significance level of .
Table 2.
The number (percentage) of non-missing observations at each time point in Figure 1.
Table 3.
The baseline characteristics of 638 subjects, including gender, age, years of education, handedness (R/L) and intracranial volume (ICV).
Table 4.
56 cortical thickness and volumetric phenotypes.
Table 5.
Significant SNPs and each one's associated phenotype numbers at the level of .
Table 6.
Significant SNPs and each one's associated phenotype numbers at the level of .
Table 7.
The numbers of the significant SNP-phenotype associations at various levels of false discovery rate (FDR).
Figure 2.
Comparison of the Manhattan plots for genome-wide p-values for phenotype left hippocampus volume from longitudinal analysis (left) and from cross-sectional analysis (right); SNP rs429358 is not included due to its small p-value.
Figure 3.
Comparison of the Q-Q plots for genome-wide p-values for phenotype left hippocampus volume from longitudinal analysis (left) and from cross-sectional analysis (right); SNP rs429358 is not included.
Figure 4.
Comparison of the Manhattan plots for genome-wide p-values for phenotype volume of right inferior lateral ventricle from longitudinal analysis (left) and cross-sectional analysis (right); SNP rs429358 is not included.
Figure 5.
Comparison of the Q-Q plots for genome-wide p-values for phenotype volume of right inferior lateral ventricle from longitudinal analysis (left) and from cross-sectional analysis (right); SNP rs429358 is not included.
Figure 6.
Comparison of the Manhattan plots without (left) or with (right) top 10 PCs.
Figure 7.
Comparison of the Q-Q plots without (left) or with (right) top 10 PCs.
Figure 8.
The Q-Q plots for genome-wide p-values for phenotype left hippocampus volume from longitudinal analysis based on (a) GEE with the sandwich covariance estimator (left, inflation factor = 1.070), (b) GEE with the model-based covariance estimator (middle, inflation factor = 2.077), and (c) linear mixed model with only a random intercept term (right, inflation factor = 1.976).
Table 8.
Simulation results at significance level with different methods for phenotypic data generated from model (2).
Table 9.
Simulation results at significance level with different methods for phenotypic data generated from model (3).