Figure 1.
The whole workflow of EW method.
(A) Edge Weight Computation; (B) Essential Protein Identification based on Edge Weights.
Figure 2.
EW performance compare with NC and PeC methods.
(A)(B): Comparison among the numbers of essential proteins identified by EW, NC and PeC on Yeast and E.coli, respectively, when selecting 50, 100, 200, 300, 400, 500 and 600 top ranked proteins; (C)(D): PR curves of EW, NC and PeC on Yeast and E.coli.
Figure 3.
PR curves of EW methods on the perturbed Yeast and E.coli PPI networks for essential protein identification.
The PPI networks are perturbed from the top, the bottom and randomly by deleting X edges in 10 steps in the ranked EW edge list. (A1, B1) From the top for X = 100 on Yeast and E. coli; (A2, B2) Form the bottom for X = 1000 and 500 on Yeast and E. coli, respectively; (A3, B3) For X = 1000 and 500 randomly on Yeast and E. coli, respectively.
Figure 4.
Sub-networks in (A) E.coli and (B) Yeast PPI network formed by top edges.
The red nodes are essential proteins and the green ones are non-essential proteins.
Table 1.
Modular examples pathway enrichment analysis by DAVID.
Figure 5.
Modular examples in Yeast and E.coli PPI network by top edges.
The red nodes are essential proteins and the green ones are non-essential proteins.