TABLE 2

Detection and characterization of imprinted QTL

Simulation detailsPowerbEstimated effectscQTL positioncImprinting inferred
No. males/females F1QTL effecta (frequency)Mend.Imp.dMend. â ± SDeImp.dâ ± SDeMend. cM ± SDImp.d cM ± SDFMend e Fred f Bothg
20/80P 0.75 (1.0/0.0)1.01.00.75 ± 0.070.75 ± 0.0646 ± 2.446 ± 1.41.00.960.96
P 0.50 (1.0/0.0)1.01.00.50 ± 0.070.50 ± 0.0546 ± 4.246 ± 2.11.00.960.96
P 0.25 (1.0/0.0)0.831.00.28 ± 0.050.25 ± 0.0546 ± 11.346 ± 6.80.970.950.92
P 0.75 (0.8/0.2)0.950.990.46 ± 0.110.45 ± 0.1146 ± 7.146 ± 4.20.980.940.93
P 0.50 (0.8/0.2)0.820.980.33 ± 0.080.31 ± 0.0846 ± 10.346 ± 6.60.940.930.90
P 0.25 (0.8/0.2)0.330.740.22 ± 0.060.18 ± 0.0447 ± 17.446 ± 12.90.610.700.59
M 0.75 (0.8/0.2)0.971.00.46 ± 0.100.45 ± 0.0947 ± 8.146 ± 3.60.990.940.94
M 0.50 (0.8/0.2)0.850.990.33 ± 0.070.30 ± 0.0646 ± 11.446 ± 6.30.970.930.91
M 0.25 (0.8/0.2)0.330.740.22 ± 0.060.18 ± 0.0447 ± 17.346 ± 13.30.590.70.56
2/80P 0.75 (0.8/0.2)0.840.850.51 ± 0.290.50 ± 0.2946 ± 6.746 ± 4.00.850.800.80
P 0.50 (0.8/0.2)0.760.840.36 ± 0.200.34 ± 0.1946 ± 9.446 ± 5.50.820.790.77
P 0.25 (0.8/0.2)0.450.690.24 ± 0.100.20 ± 0.1047 ± 15.146 ± 11.00.600.640.56
M 0.75 (0.8/0.2)0.991.00.46 ± 0.100.45 ± 0.0846 ± 7.046 ± 3.51.00.950.95
M 0.50 (0.8/0.2)0.880.990.32 ± 0.070.30 ± 0.0647 ± 10.446 ± 5.80.970.940.93
M 0.25 (0.8/0.2)0.370.780.22 ± 0.060.18 ± 0.0448 ± 19.246 ± 13.40.640.740.62

QTL were simulated under an imprinting model and analyzed under Mendelian (Mend.) and imprinting (Imp.) models for 400 F2 individuals with different designs, QTL effects, and allele frequencies.

  • a P, paternally expressed QTL effect; M, maternally expressed QTL effect (frequency of positive QTL allele in F0).

  • b Proportion of replicates significant at the 5% chromosome-wide level against the H0 of no QTL.

  • c Estimates and empirical standard deviations, calculated with the replicates that exceed the 5% chromosome-wide significance level.

  • d Analyzed under the appropriate reduced model (Equation 12).

  • e Proportion of replicates significant at the 5% chromosome-wide level and for which a full model explains significantly more variance (P < 0.05) than a Mendelian QTL at the position of the best QTL under the respective model.

  • f Proportion of replicates significant at the 5% chromosome-wide level, for which a full model does not explain significantly more variance (P < 0.05) than a QTL with a single parental effect at the position of the best imprinted QTL.

  • g Proportion of replicates where tests of both full vs. Mendelian (FMend) and reduced vs. full (Fred) indicate imprinting.