A knowledge-guided strategy for improving the accuracy of scoring functions in binding affinity prediction

BMC Bioinformatics

Table 2 Statistical results produced by two scoring functions alone on the three test sets

	X-Score			PLP
Test set	R ^a	SD ^b	k ^c	R ^a	SD ^b	k ^c
HIV protease complexes (N = 112)	0.329	1.55	0.664	0.190	1.61	-0.0099
Carbonic anhydrase complexes (N = 44)	0.648	1.06	2.045	0.690	1.01	-0.0833
Trypsin complexes (N = 73)	0.815	0.98	1.988	0.762	1.09	-0.0518

^aPearson correlation coefficient between the experimental binding constants of these complexes and the binding scores produced by the given scoring function.
^bStandard deviation in regression (in logK_dunits)
^cSlope of the regression line, which is needed in Equation 4.

ISSN: 1471-2105