Skip to main content

Table 2 Comparison of protein complex prediction by SGD reference dataset

From: Finding low-conductance sets with dense interactions (FLCD) for better protein complex prediction

Network Method # complex #. matched coverage Recall Precision F-measure Sn PPV Acc MMR
SceDIP FLCD 2134 152 3921 0.6786 0.2020 0.3113 0.5964 0.5003 0.5462 0.3685
  CONE 380 86 1503 0.3839 0.2579 0.3085 0.4082 0.6203 0.5032 0.1950
  LinkC 1839 137 3735 0.6116 0.1289 0.2130 0.6290 0.4820 0.5506 0.3276
  SR-MCL 3216 44 4678 0.2228 0.0221 0.0412 0.5120 0.2893 0.3489 0.0708
SceBG FLCD 4027 183 5836 0.7821 0.2000 0.3181 0.7363 0.5621 0.6433 0.4920
  CONE 522 122 2735 0.5214 0.2433 0.3318 0.6488 0.6035 0.6257 0.2542
  LinkC 5382 164 6076 0.7008 0.1217 0.2072 0.8880 0.4373 0.6231 0.4100
  SR-MCL 1862 108 5889 0.4615 0.1245 0.1961 0.8999 0.3034 0.5225 0.2151
SceIntAct FLCD 3394 172 4678 0.7446 0.1933 0.3069 0.6699 0.5391 0.6009 0.4661
  CONE 496 117 1994 0.5065 0.2419 0.3275 0.5742 0.5944 0.5842 0.2742
  LinkC 1297 93 5290 0.4026 0.0941 0.1525 0.9223 0.2393 0.4698 0.2285
  SR-MCL 1079 68 5342 0.2294 0.0437 0.1517 0.7784 0.2402 0.4341 0.1213
SceMINT FLCD 2483 157 4210 0.6826 0.2280 0.3418 0.6524 0.5284 0.5871 0.4163
  CONE 513 110 2335 0.4783 0.2027 0.2848 0.5370 0.5954 0.5654 0.2442
  LinkC 2201 144 4068 0.6261 0.1595 0.2542 0.6757 0.5540 0.6119 0.3743
  SR-MCL 3698 33 4976 0.1435 0.0169 0.0302 0.5013 0.2597 0.3608 0.0609
  1. CONE and LinkC are short for ClusterONE and LinkComm, respectively
  2. Bold values denote the best scores corresponding to specific criteria