Skip to main content

Table 5 The effect of network size on GRN reconstruction in case of non-linearity between the genes. Results are the average of the accuracy in terms of the Link criteria for GRN reconstruction of 10 different randomly generated synthetic networks

From: Identifying time-delayed gene regulatory networks via an evolvable hierarchical recurrent neural network

Methods P TP FP FN Precision Recall F-score Nominal p-value Adjusted p-value
HRNN 5 5.6 3.4 4.7 0.62 0.54 0.57   
  10 6.4 12.8 12.6 0.33 0.33 0.33   
  20 11.2 30.5 28.4 0.26 0.28 0.27   
  30 10.9 50.3 46.1 0.18 0.19 0.18   
TD-ARACNE 5 6.60 6.9 3.7 0.50 0.64 0.54 0.60 1
  10 9.3 35.7 9.7 0.21 0.50 0.28 0.16 0.64
  20 14.5 129.5 25.1 0.10 0.37 0.15 3.0×10−4 1.2×10−3
  30 12 162 45 0.07 0.20 0.09 2.0×10−5 8.0×10−5
HCC-CLINDE 5 3.1 1.8 7.2 0.65 0.29 0.39 7.0×10−4 2.8×10−3
  10 3.3 4.4 15.7 0.46 0.17 0.24 0.05 0.2
  20 4.1 9.4 35.5 0.29 0.10 0.15 0.007 0.028
  30 3.4 13.0 53.6 0.21 0.05 0.09 0.002 0.008
TSNI 5 10.3 14.7 0.0 0.41 1 0.58 0.91 1
  10 16.9 71.1 2.1 0.19 0.89 0.31 0.60 1
  20 17.6 160.4 22.0 0.09 0.44 0.16 4.0×10−4 1.6×10−3
  30 10 158 47 0.06 0.17 0.08 2.0×10−4 8.0×10−4
ebdbNet 5 4.4 6.1 5.9 0.43 0.43 0.41 0.008 0.032
  10 4.8 20.2 14.2 0.19 0.25 0.20 0.004 0.016
  20 6.2 55.80 33.4 0.10 0.15 0.12 5.0×10−5 2.0×10−4
  30 3.8 59.2 53.2 0.06 0.06 0.06 7.9×10−5 3.1×10−4
  1. The p-values are for how the Link F-scores of other methods compare with HRNN. P is the number of genes for each of the networks. The variance of the noise is equal to 1