Identification of binding pockets in protein structures using a knowledge-based potential derived from local structural similarities

Table 3 Predicted binding residues for the GDPRAN-NTF2 complex

	CHAIN	RESIDUE	RESIDUE NUM	BIND	NON-BIND	PDBINDER PROPENSITY VALUE
PDBinder predictions	D	G	19	311	627	0.332
	D	G	20	409	487	0.456
	D	T	21	200	491	0.289
	D	G	22	868	754	0.535
	D	K	23	1112	267	0.806
	D	T	24	476	557	0.461
	D	T	25	418	535	0.439
	D	N	122	405	244	0.624
	D	K	123	381	156	0.709
	D	D	125	145	163	0.471
	D	S	150	227	433	0.344
	D	A	151	343	809	0.298
	D	K	152	31	175	0.150
Q-SiteFinder predictions	A	Q	65	1	109	0.009
	A	S	67	57	532	0.097
	A	Q	88	14	221	0.060
	A	L	89	24	1168	0.020
	A	I	96	9	313	0.028
	A	G	98	82	741	0.100
	A	A	122	16	538	0.029
	A	H	124	3	26	0.103
	B	M	97	8	207	0.037
	B	G	98	96	741	0.115
	B	A	122	19	568	0.032
	B	H	124	3	22	0.120
	B	F	126	0	10	0.000
	D	V	40	33	1820	0.018
	D	A	41	44	1033	0.041
	D	T	42	36	426	0.078

Binding site residues predicted by Q-SiteFinder and PDBinder on GDPRAN-NTF2 complex (PDB: 1a2k). The only false positive residue identified by PDBinder is S150.

ISSN: 1471-2105