Skip to main content
. 2019 Apr 2;9:5479. doi: 10.1038/s41598-019-41918-3

Figure 2.

Figure 2

The average differential networks estimated by each method based on 20 generated samples of size n=250. The cells are shaded based on how frequently the edge was identified as differentially connected in the 20 samples; darker cells indicate higher frequency. The true differential network is based on the absolute difference in the precision matrices of the two populations.