Table 4.
Number of genes covered by reverse-engineered networks. Shows the sizes of networks constructed from complete dataset when (A) IQR threshold q = 0.65 and (B) q is selected from the histogram plot. Also shown in the table are the sizes of networks constructed by [7,8,9].
| Network | Size of Input Datset | Genes in Input Dataset | Genes in Network |
|---|---|---|---|
| PCC with dataset (A) | 11,760 | 13,384 | 2670 |
| MI with dataset (A) | 11,760 | 13,384 | 13,181 |
| PCC with dataset (B) | 11,760 | 18,806 | 3940 |
| MI with dataset (B) | 11,760 | 18,806 | 18,606 |
| MI network by [7] | 3137 | 15,578 | 15,495 |
| PCC network by [8] | 1094 | 16,293 | 6206 |
| GGM network by [9] | 2045 | NA 1 | 6760 |
1 [9] constructs the GGM network by randomly sampling 2000 genes at a time. NA 1: Not Applicable; PCC: Pearson correlation coefficient; MI: Mutual Information; GGM: Gaussian Graphical Model.