Page 115 - 76_02

VOL. 76 (2), 223-239, 2010 PREDICTION OF LIGAND BINDING ENERGY...

properly described, so lower values of b should be required in some
cases.

Table 4. Test set energies and LIE parameters.

Compound Free Evdw Free Eqq OH ?Gcalc
(KJ/mol) (KJ/mol)
A B Groups (kcal/mol)

1 12.727 78.981 0.42 0.52 0 –10.73
79.792 0.42 0.52 1 –11.21
2 12.817 79.787 0.42 0.52 0 –11.21
82.810 0.42 0.52 1 –11.64
3 12.830 81.671 0.42 0.52 0 –11.48
84.418 0.42 0.52 0 –11.87
4 13.328 82.770 0.42 0.52 0 –11.24
79.895 0.42 0.52 1 –11.23
5 13.147 78.919 0.42 0.52 0 –10.72
84.461 0.42 0.52 1 –11.47
6 13.586 79.876 0.42 0.50 2 –11.41
82.791 0.42 0.50 3 –10.84
7 13.327 85.701 0.42 0.50 3 –11.63
85.101 0.42 0.50 2 –11.55
8 12.874 85.188 0.42 0.48 4 –10.75
85.706 0.42 0.50 3 –11,57
9 12.710 84.266 0.42 0.50 3 –11.38
86.480 0.42 0.50 3 –11.71
10 13.599

11 12.846

12 13.299

13 13.762

14 13.683

15 13.686

Dexamethasone 13.805

Betamethasone 13.578

Cortisol 13.874

Figure 4 shows the regression test performed between the values
of calculated energy of binding and the experimental values of
EC150. The squared correlation coefficient was found significant
(R2 = 0.74 and p = 0.000), hence a good correlation between
calculated energy values and the agonist effect of the compounds
can be expected.

However, although an acceptable value of correlation was found,
in line with similar recent studies (16), some authors found slight

235