Table 1.

Prospective test: Absolute binding free energy predictions

	Structure	ΔGb,exp (Kcal/mol) a	ΔG b,calc (Kcal/mol) b	RMSD (Å) c	PDB ID
n-phenylglycinonitrile		−5.52 ± 0.18	−5.63 ± 0.38	0.87	2RBO
2-nitrothiophene		−4.85 ± 0.25	−5.73 ± 0.13	1.09(A); 2.86(B)	2RBN
thieno[3,2-b]thiophene		−4.67 ± 0.17	−6.8 ± 0.23	0.73(A) 0.44(B)	3HUQ
4,5,6,7-tetrahydroindole		−4.61 ± 0.09	−5.4 ± 0.45	0.66 & 1.78**	3HUA
benzyl acetate		−4.48 ± 0.16	−1.31 ± 0.44	>10	3HUK
nitrosobenzene		weak	−5.55 ± 0.23	3.24(A), 3.22(B)	3HU9
2-ethoxy-3,4-dihydro-2h-pyran		weak	−3.8 ± 0.18	4.35	3HTG
4-chloro-1h-pyrazole		weak	−7.86 ± 0.12	2.07	3HTF
(E)-thiophene-2-carboxaldoxime		NB	−2.30 ± 0.09	---	---
1-phenylsemicarbazide		NB	0.45 ± 0.24	---	---
o-benzylhydroxylamine		NB	−2.58 ± 0.13	---	---
1-2-hydroxyethylpyrrole		NB	−5.72 ± 0.12	---	---
phenylhydrazine		NB	−2.51 ± 0.44	---	---

Compound names of mispredictions are in bold italics.

^AFree energy of binding determined by ITC at 10°C. For ligands designated weak binding was established by Tm upshift but ΔG_b could not be determined; NB = nonbinder, ΔTm ≈ 0°C at concentrations between 1–10mM.

^BCalculated free energy of binding.

^CRMSD of predicted ligand geometry to experimentally observed crystal pose. If multiple ligand orientations were present in the crystal (designated A and B) only the best RMSD to the prediction is reported.

^**indicates RMSD for one crystal pose calculated to two predictions.