Table I Comparison of Network estimation of the optimal models in all the three simulations
ModelsFNFPTFFDR
Simulation I
    ProMAP.comp43.8 (14.32)16.4 (6.72)60.2 (12.7)0.06 (0.02)
    remMap.comp75.75 (17.76)13.55 (5.24)89.3 (16.81)0.05 (0.02)
    corr.test.comp87.65 (21.81)15.75 (6.18)103.4 (19.28)0.07 (0.02)
    ProMAP.obs65.55 (18.31)28.85 (16.5)94.4 (17.51)0.1 (0.05)
    remMap.obs43.3 (12.57)68.15 (14.87)111.45 (13.34)0.2 (0.03)
    corr.test.obs95.2 (20.37)38.35 (10.47)133.55 (15.02)0.15 (0.03)
Simulation II
    ProMAP.comp166.55 (23.03)17.3 (5.28)183.85 (20.61)0.04 (0.01)
    remMap.comp184.3 (26.42)42.25 (10.67)226.55 (26.47)0.09 (0.02)
    corr.test.comp356.4 (13.84)2.3 (1.75)358.7 (13.55)0.01 (0.01)
    ProMAP.obs219.15 (24.57)56.6 (9.08)275.75 (22.13)0.13 (0.02)
    remMap.obs214.5 (24.36)80.4 (13.53)294.9 (22.08)0.17 (0.02)
    corr.test.obs356.25 (12.85)29.3 (7.05)385.55 (11.66)0.11 (0.02)
Simulation III
    ProMAP.comp226.3 (21.05)83.5 (14.68)309.8 (28.37)0.18 (0.03)
    remMap.comp265.3 (25.03)82.65 (17.32)347.95 (31.7)0.2 (0.04)
    corr.test.comp303.95 (12.98)489.1 (35.96)793.05 (31.59)0.62 (0.02)
    ProMAP.obs295.25 (21.69)84.15 (22.19)379.4 (28.36)0.21 (0.04)
    remMap.obs313.85 (23.67)76.85 (14.85)390.7 (30.07)0.21 (0.04)
    corr.test.obs304.8 (13.08)639.6 (46.69)944.4 (40.13)0.68 (0.01)
  • Network topology: p = 80, L = 100 for Simulation I with 310 edges; p = 1000, L = 1200 for Simulation II & III with 600 edges. FP: false positive; FN: false negative; TF: total false; FDR: false discovery rate. Numbers in parentheses are standard deviations over 20 replicates.