Skip to main content

Table 1 Performances (mean% ± std.% of AUROCs, mean% ± std.% of AUPRs) of different inference methods on Simulation data I (n = 20, SNR = 10, 20 and 30). Best results for each SNR case are in bold

From: Globally learning gene regulatory networks based on hidden atomic regulators from transcriptomic big data

METHOD SNR = 10 SNR = 20 SNR = 30
GENIE3 81.13 ± 0.49, 53.83 ± 0.77 82.05 ± 0.51, 56.11 ± 0.61 81.89 ± 0.51, 56.04 ± 0.78
CLR 81.04 ± 0.29, 57.66 ± 0.42 81.95 ± 0.36, 59.80 ± 0.53 81.74 ± 0.44, 59.50 ± 0.69
ARACNe-AP 62.56 ± 1.45, 13.68 ± 5.37 64.17 ± 0.61, 18.63 ± 3.77 64.13 ± 1.32, 19.17 ± 7.03
ARACNE 81.84 ± 0.43, 55.20 ± 0.53 82.65 ± 0.41, 57.18 ± 0.73 82.55 ± 0.62, 56.37 ± 1.10
dlGRN (l = 25) 88.21 ± 0.51, 65.49 ± 1.10 90.05 ± 0.67, 70.10 ± 1.49 90.62 ± 0.41, 71.05 ± 0.89
dlGRN (l = 50) 91.06 ± 0.31, 77.77 ± 1.00 96.10 ± 0.42, 89.88 ± 1.12 97.45 ± 0.29, 92.11 ± 1.48
dlGRN (l = 100) 90.48 ± 0.45, 75.16 ± 1.06 96.23 ± 0.65, 89.18 ± 1.39 97.73 ± 0.32, 92.06 ± 0.22
dlGRN (l = 150) 90.40 ± 0.45, 75.86 ± 0.98 95.96 ± 0.32, 87.73 ± 0.41 97.54 ± 0.46, 91.44 ± 1.13