Skip to main content

Table 2 Number of predicting high and low connectivity essential proteins by using SON and other nine existing methods

From: Predicting essential proteins based on subcellular localization, orthology and PPI networks

  K DC IC EC SC BC CC NC PeC ION SON
degree < =10 1 % 0 0 0 0 0 0 0 0 17 14
5 % 0 0 0 0 0 0 3 40 66 64
10 % 0 0 0 0 1 0 27 84 108 116
15 % 0 0 8 8 18 7 66 117 146 156
20 % 0 0 28 28 41 20 101 153 188 193
25 % 11 20 73 73 76 55 156 192 253 220
degree > 10 1 % 22 24 24 24 24 24 32 39 24 29
5 % 101 102 96 96 95 104 156 133 122 129
10 % 207 210 195 195 181 193 255 209 223 233
15 % 320 316 271 271 253 277 307 255 299 313
20 % 413 406 349 349 320 344 363 285 359 387
25 % 491 484 394 394 357 393 388 302 381 441
  1. As shown in the top part of Table 2 (degree < = 10), it is weak for eight centrality methods to predict low connectivity essential proteins. When taking the top 20 % proteins ranked in descending order according to their ranking scores computed by DC and IC, the numbers of predicting essential proteins are 0. The performance of SON overall is better than that of eight centrality methods (DC, IC, EC, SC, BC, CC, NC and PeC). When K is 10, 15, 20 %, respectively, the performance of SON is also better than that of ION