3D Pharmacophore, hierarchical methods, and 5-HT₄ receptor binding data

Figures & data

Figure 1 General representation of the pharmacomodulation program.

Figure 2 5-HT₃ partial agonist pharmacophore.

Figure 3 5-HT₄ antagonist pharmacophore.

Scheme 1 Structures of compounds 1 and 2.

Figure 4 Frequencies vs Class (%/100) associated with the percentage of inhibition at 10^{− 6}M.

Table I.  Dataset dissimilarity values.

Dataset/Dataset	ANN	AFN	OAN
Overall/Overall	0.10	0.87	0.68
Actives/Inactives	0.42	0.80	0.67
Actives/Actives	0.06	0.76	0.46
Inactives/Inactives	0.12	0.87	0.70

Figure 5 Overall Density Neighbors density representation. (black: overall; green: actives-inactives; red: actives-actives; brown: inactives-inactives).

Figure 6 Evolution of the different indexes as a function of clustering level (Unity fingerprint/Tanimoto/Energy). (Pa curve in red, QCI curve in green; QCI_W curve in blue).

Table II.  Classification of the hierarchies (the first twenty) as a function of QCI_WD values.

Fingerprint	Metric	Method	QCIWD
Pfp_JChem	Canberra	Energy	11.77
Pfp_JChem	Canberra	Ward	11.67
Pfp_JChem	Binary	Ward	11.54
Ufp_Unity	Tanimoto	Energy	11.35
Ufp_Unity	MT	Energy	11.33
Pfp_JChem	Binary	Energy	11.31
Ufp_Unity	Dice	Energy	11.30
Ufp_Unity	Euclidean	Energy	11.29
Ufp_Unity	MT	Ward	11.22
Cfp_JChem	MT	Energy	11.17
Pfp_JChem	Canberra	Complete	11.17
Cfp_JChem	Euclidean	Energy	11.17
Ufp_Unity	Tanimoto	Ward	11.17
Cfp_JChem	Euclidean	Ward	11.17
Fpfp_Jchem	Euclidean	Energy	11.14
Pfp_JChem	Euclidean	Energy	11.14
Cfp_JChem	Tanimoto	Energy	11.13
Ufp_Unity	Euclidean	Ward	11.12
Cfp_JChem	MT	Ward	11.08
Ufp_Unity	Dice	Ward	11.07

Table III.  Comparison as a function of QCI_WD values.

Fingerprint	Metric	Energy	Ward	Complete	Average	Single
Cfp_Jchem	Euclidean	11.17	11.17	10.60	10.26	9.40
Cfp_Jchem	Tanimoto	11.13	10.97	9.99	9.76	8.86
Cfp_Jchem	Dice	11.03	10.90	9.99	9.78	8.86
Cfp_Jchem	MT	11.17	11.08	10.39	9.99	9.07
Ufp_Unity	Euclidean	11.29	11.12	10.33	9.97	9.04
Ufp_Unity	Tanimoto	11.35	11.17	9.99	9.73	8.77
Ufp_Unity	Dice	11.30	11.07	9.99	9.73	8.77
Ufp_Unity	MT	11.33	11.22	9.94	9.77	8.83
Pfp_Jchem	Euclidean	11.14	11.05	10.50	10.23	6.18
Pfp_Jchem	Maximum	10.16	9.83	9.40	9.21	6.01
Pfp_Jchem	Canberra	11.77	11.67	11.17	10.68	7.44
Pfp_Jchem	Binary	11.31	11.54	10.62	10.20	7.33
Fpfp_Jchem	Euclidean	11.14	10.96	10.34	9.58	7.55
Fpfp_Jchem	Maximum	10.50	10.35	10.10	9.34	7.64
Fpfp_Jchem	Canberra	10.78	10.69	9.87	9.16	6.90
Fpfp_Jchem	Binary	8.78	7.88	7.42	6.69	5.77
Average		10.96	10.79	10.04	9.63	7.90
Standard Deviation		0.69	0.89	0.81	0.88	1.21

Figure 7 Dendrogram associated corresponding to the best hierarchy based on pharmacophore fingerprint (Pfp_Jchem/Canberra/Energy). The numbers indicate the branches corresponding to active clusters (on the right of the number).

Figure 8 Dendrogram corresponding to the best hierarchy based on chemical fingerprint (Ufp_Unity/Tanimoto/Energy).

Table IV.  Line up of active and inactive clusters as a function of the activity and fit values.

			Number of compounds
Activity of clusters	Activity of compounds	Fit the pharmacophore	Pfp/Canberra/Energy	Unity/Tanimoto/Energy
Active	Active	Yes	114	103
		No	29	24
	Inactive	Yes	30	24
		No	27	41
Inactive	Active	Yes	5	16
		No	15	20
	Inactive	Yes	97	103
		No	683	669

Table V.  Percentage of compounds in active and inactive clusters in relation to fit towards the 3D pharmacophore.

			Percentage of compounds
Activity of clusters	Fit the pharmacophore	Compounds	Pfp/Canberra/Energy	Unity/Tanimoto/Energy
Active clusters	Yes	All	72 (+47)	66 (+41)
		Active	80 (+8)	81 (+9)
		Inactive	53 (+28)	37 (+22)
Inactive clusters	No	All	87 (+12)	85 (+10)
		Inactive	88 (+13)	87 (+12)
		Active	75 (+50)	56 (+31)