Searching QTL for several traits with many markers as in the preliminary experiment increases the chances of type I error (i.e., a false positive association).
Weller et al. (1998) proposed applying the false discovery rate (FDR) for multiple comparisons in QTL analysis. In the current preliminary experiment, there were 40 marker–trait combinations (two traits, 20 markers), of which 5 were significant at P < 0.028 (Table 1).
The critical P-value of 2.8% implies that 1 out of the 40 combinations is expected asfalse positive.
Therefore, four of the five associations found should be true effects (FDR = 20%).
The second experiment conducted with a larger test population and only six markers ensured better protection against false discoveries (Weller et al., 1998).
Searching QTL for several traits with many markers as in the preliminary experiment increases the chances of type I error (i.e., a false positive association). Weller et al. (1998) proposed applying the false discovery rate (FDR) for multiple comparisons in QTL analysis. In the current preliminary experiment, there were 40 marker–trait combinations (two traits, 20 markers), of which 5 were significant at P < 0.028 (Table 1). The critical P-value of 2.8% implies that 1 out of the 40 combinations is expected asfalse positive. Therefore, four of the five associations found should be true effects (FDR = 20%). The second experiment conducted with a larger test population and only six markers ensured better protection against false discoveries (Weller et al., 1998).
การแปล กรุณารอสักครู่..