Statistical deconvolution of enthalpic energetic contributions to MHC-peptide binding affinity

BMC Structural Biology

Table 5 Best QSAR models for GFA and G/PLS based on binding Pockets in comparison with the PLS method. * PLS Models based on Cerius GFA & G/PLS Models respectively.

Model	QSAR Analysis¹Cerius²SYBYL	Equation Length Terms + Constant	STATISTICAL PARAMETERS
			TRAINING SET									TEST SET
			LOF	SEP	r ²		(CV) r ²		SEE	PRESS		r		r ²		PRESS
Total + Electrostatic + VDW	GFA ¹	13	0.204		0.796		0.681			12.949		0.762		0.499		41.896
	G/PLS ¹	16	-		0.789		0.611			16.029		0.843		0.711		37.423
	PLS ² (GFA)	13		0.593		0.546		0.439	0.533		19.349		0.714		0.510		41.707
	PLS ² (G/PLS)	16		0.704		0.432		0.284	0.628		30.279		0.818		0.699		42.235
Electrostatic + VDW	GFA ¹	13	0.242		0.751		0.647			15.194		0.698		0.487		37.794
	G/PLS ¹	19	-		0.643		0.355			23.922		0.605		0.366		40.364
	PLS ² (GFA)	13		0.548		0.659		0.528	0.466		14.709		0.344		0.118		51.406
	PLS ² (G/PLS)	19		0.729		0.523		0.254	0.583		25.187		0.480		0.230		45.155
Total	GFA ¹	20	0.280		0.810		0.624			-		0.210		0.044		60.931
	G/PLS ¹	20	-		0.528		0.072			28.473		0.513		0.263		58.726
	PLS ² (GFA)	20		0.627		0.539		0.369	0.536		19.527		0.636		0.405		47.833
	PLS ² (G/PLS)	20		0.654		0.434		0.256	0.571		24.467		0.625		0.425		52.133

ISSN: 1472-6807