TABLE 4

Comparison-wise error rate (CWER) thresholds to control genome-wise type I error rates (αG) or false discovery rates (αF) at 10, 5, or 1% for five traits separately or across the five traits based on four approaches

αG = αF = 10%αG = αF = 5%αG = αF = 1%
TraitLKCDFDRmFDR1LKCDFDRmFDR1LKCDFDRmFDR1
Carcass wt.0.0000890.000680.00433a0.005810.0000340.000310.001100.001800.0000060.0000550.0000100.000081
0.00900b0.003400.000039
Back fat0.0000890.000620.007070.008600.0000340.000280.001000.002180.0000060.000050
0.016660.00256
Loin eye area0.0000890.000680.001630.001480.0000340.000310.000420.000190.0000060.0000600.000067
0.003060.001000.000160
Cholesterol0.0000890.000620.0000340.000280.0000060.000055
Marbling0.0000890.000680.000680.0000340.000340.0000060.000067
0.00170
Across traits0.0000140.000120.001480.002410.0000070.000060.000260.000140.0000010.000011
0.003890.00045

LK, Lander and Kruglyak approach; CD, Churchill and Doerge permutation test; FDRm, false discovery rate based on the highest test per marker interval; FDR1, false discovery rate based on tests at every 1-cM position.

  • a Comparison-wise error rate (CWER) for the maximum test in the interval, based on the standard F-distribution for a single test, ignoring the effective number of tests in the interval.

  • b Interval-wise error rate (IWER) for the maximum test in the interval, corrected for effective number of tests in the interval.