Skip to main content

Table 2 Type I error estimate at alpha = 0.05 with 10,000 replicates of phenotype and a single SNP in LE with a QTL explaining 10% phenotype variation.

From: A three-stage approach for genome-wide association studies with family data for quantitative traits

MAF Polygenic Variance LM LME FBAT LM-LME 1 LM-LME 2
0.005 0.3 0.087 0.046 0.052 0.044 0.045
0.005 0.6 0.118 0.053 0.047 0.046 0.050
0.01 0.3 0.089 0.050 0.049 0.048 0.050
0.01 0.6 0.122 0.049 0.048 0.044 0.048
0.05 0.3 0.087 0.048 0.052 0.047 0.048
0.05 0.6 0.118 0.050 0.050 0.043 0.047
0.1 0.3 0.082 0.044 0.053 0.041 0.043
0.1 0.6 0.122 0.049 0.054 0.043 0.047
  1. The SNP explains 0% phenotype variation.
  2. A QTL is simulated to explain 10% phenotype variation and in LE with the SNP.
  3. Marginal phenotype distribution follows the standard normal distribution.
  4. LM_LME1: two-stage approach with LM (p-value cut-off 0.1) at first stage
  5. LM_LME2: two-stage approach with LM (p-value cut-off 0.2) at first stage