Skip to main content

Advertisement

Table 2 FEATURE benchmark of the 20 apo-holo pairs of calcium-binding proteins

From: Prediction of calcium-binding sites by combining loop-modeling with machine learning

Protein and calcium ID Holo structure Apo structure Apo-gap Apo-loop
1B9AA110 81.72 26.75 5.41 51.58
1C1R_247 56.57 51.71 1.94 55.43
1DPO_246 52.89 47.94 -0.17 50.87
1DVIB275 66.89 10.63 0.00 29.70
1DVIB277 68.39 57.51 -18.90 41.18
1F6SB202 68.11 54.04 16.58 54.46
1HAZB1246 51.82 48.21 -4.82 32.72
1I40A305 30.52 27.87 16.92 46.64
1K94_998 80.71 63.46 -22.81 42.92
1K94_999 40.31 32.63 -0.45 50.96
1K96A91 79.61 45.63 9.77 59.27
1KXQB4003 54.99 43.25 22.01 59.53
1NLS_240 61.98 52.09 28.46 60.76
1NOL_632 36.88 30.22 30.22 41.85
1PSH_1 34.09 33.81 22.03 56.37
1QMDA403 60.45 30.41 27.37 50.41
1QMDA405 59.81 45.00 37.05 52.09
2POR_304 47.41 71.34 34.96 89.29
3LHM_131 61.69 54.34 2.46 58.95
5CHY_401 50.19 46.70 23.56 54.08
  1. We compare the performance of FEATURE on the holo structures, the apo structures, the apo structures of which the binding loops were removed (apo-gap), and the apo structures of which the binding loops (apo-loop) have been rebuilt using modeling methods. Protein and calcium ID are from PDB (column 1). FEATURE scores that evaluate the probability of the likelihood of a site being calcium-binging are listed in column 2-5. Of the 20 pairs, 17 apo-loops have higher FEATURE scores than the corresponding apo structures. Bolded number indicates that FEATURE identifies the calcium-binding site correctly using a score threshold of 50. FEATURE identifies 15 out of 20 calcium-binding sites in the holo structures. Combined with modeling methods, FEATURE identifies 14 out of 20 calcium-binding sites in the apo-loops. For comparison, FEATURE identifies binding sites correctly in only 7 out of 20 the apo structures, and 0 out of 20 the apo-gaps. These results demonstrate that reconstruction of the calcium-binding loops in the apo structures allows FEATURE to identify cryptic calcium sites effectively.