Topological descriptors in modeling malonyl coenzyme A decarboxylase inhibitory activity: N-Alkyl-N-(1,1,1,3,3,3-hexafluoro-2-hydroxypropylphenyl)amide derivatives

Figures & data

Table I.  Observed and modeled malonyl coenzyme A decarboxylase inhibition activity of N-alkyl-N-(1,1,1,3,3,3-hexafluoro-2-hydroxypropylphenyl)amide derivatives.

						–logIC50 (M)
S.No	R1	R2	CIC2	MATS6e	GATS8e	Obsd^*	Calcd Eq (3)
1	Me	CH2OPh	1.704	− 0.223	0.758	6.03	5.87
2	H	CH2OPh	1.470	− 0.310	0.982	5.18	5.19
3	H	CH2CH2Ph	1.638	− 0.227	0.913	5.36	5.54
4	Me	CH2CH2Ph	1.851	− 0.127	0.731	6.11	6.19
5	Et	CH2CH2Ph	1.574	− 0.066	0.682	6.76	6.96
6	i-Pr	CH2CH2Ph	1.692	− 0.029	0.642	6.70	7.08
7	n-Bu	CH2CH2Ph	1.732	− 0.089	0.519	7.10	7.07
8	CH2COOMe	CH2CH2Ph	1.617	− 0.194	0.608	6.98	6.51
9	CH2COOH	CH2CH2Ph	1.517	− 0.206	0.716	6.43	6.30
10	benzyl	CH2CH2Ph	2.070	− .114	0.527	6.24	6.50
11	cyclohexyl	CH2CH2Ph	2.084	− 0.073	0.491	6.82	6.77
12	H	i-Pr	1.343	− 0.250	1.009	NA^†	5.56
13	Me	i-Pr	1.544	− 0.040	0.913	6.75	6.51
14	Me	3-Py	1.369	− 0.166	0.869	5.99	6.28
15	Me	4-Py	1.475	− 0.114	0.891	6.41	6.32
16	Et	i-Pr	1.513	− 0.002	0.897	7.39	6.77
17	n-Pr	i-Pr	1.455	− 0.036	0.626	7.55	7.39
18	n-Bu	i-Pr	1.490	− 0.038	0.530	7.70	7.59
19	CH2CH2NMe2	i-Pr	1.719	− 0.047	0.498	5.94^‡	7.34
20	CH2CH2OH	i-Pr	1.328	− 0.054	0.773	6.98	7.09
21	CH2COOMe	i-Pr	1.222	− 0.123	0.612	7.47	7.33
22	CH2COOH	i-Pr	1.244	− 0.128	0.774	6.20	6.86
23	(CH2)4COOMe	i-Pr	1.463	− 0.015	0.554	7.59	7.67
24	(CH2)4COOH	i-Pr	1.331	− 0.022	0.555	8.10	7.80
25	(CH2)4-(5-1H-tetrazole)	i-Pr	1.218	− 0.028	0.542	8.10	7.95
26	(CH2)4CN	i-Pr	1.344	− 0.035	0.536	7.70	7.77
27	Et	p-CNPh	1.402	− 0.067	0.701	7.43	7.12
28	Et	4-Py	1.214	− 0.055	0.791	7.02	7.19
29	H	NHPh	1.758	− 0.345	0.862	NA^†	4.97
30	Me	NHPh	1.836	− 0.225	0.705	5.81	5.83
31	H	morpholinyl	1.269	− 0.264	0.836	NA^†	6.04
32	Et	NMe2	1.587	− 0.025	0.890	7.23	6.59
33	Et	N(Et)(CH2)4OH	1.444	− 0.035	0.542	7.22	7.63
34	Me	morpholinyl	1.456	− 0.072	0.790	6.49	6.80
35	Et	morpholinyl	1.373	− 0.036	0.726	7.13	7.24
36	n-Pr	morpholinyl	1.329	− 0.079	0.578	7.47	7.48
37	n-Bu	morpholinyl	1.368	− 0.076	0.527	7.52	7.58
38	n-pentyl	morpholinyl	1.431	− 0.076	0.574	7.72	7.38
39	n-hexyl	morpholinyl	1.540	− 0.063	0.568	7.51	7.31
40	n-heptyl	morpholinyl	1.662	− 0.055	0.573	7.14	7.18
41	octyl	morpholinyl	1.790	− 0.048	0.562	6.57	7.07
42	CH2CH2Ph	morpholinyl	1.642	− 0.055	0.588	7.37	7.16
43	CH2CH2OH	morpholinyl	1.506	− 0.084	0.729	6.78	6.84
44	CH2COOMe	morpholinyl	1.329	− 0.135	0.707	7.10	6.89
45	CH2CH2COOMe	morpholinyl	1.385	0.012	0.893	7.00	7.01
46	CH2CH2COOH	morpholinyl	1.313	0.012	0.874	6.61	7.15
47	(CH2)3COOMe	morpholinyl	1.457	− 0.048	0.677	7.13	7.20
48	(CH2)3CN	morpholinyl	1.327	− 0.072	0.589	7.40	7.49
49	(CH2)4COOMe	morpholinyl	1.536	− 0.050	0.530	7.38	7.47
50	(CH2)4COOH	morpholinyl	1.417	− 0.061	0.528	7.85	7.58

* IC₅₀ represents the concentration of an inhibitor to bring out 50% inhibition of the MCD enzyme; taken from ref. (6) ^† inactive at 10 μM. ^‡ ‘outlier’ compound of present study.

Table II.  Descriptor classes used for the analysis of malonyl coenzyme A decarboxylase inhibitiory activity of N-alkyl-N-(1,1,1,3,3,3-hexafluoro-2-hydroxypropyl phenyl)amide derivatives and identified categories in modeling the activity.

Descriptor class (acronyms)	Definition and scope	Descriptors per model (Number of models)^*
Topological (TOPO)	2D-descriptor from molecular graphs & independent of conformations	2 (41)
Modified Burden eigenvalues (BCUT)	2D-descriptors representing positive & negative eigenvalues of the adjacency matrix, weights the diagonal elements and atoms	2 (1)
Galvez topological charge indices (GVZ)	2D-descriptors representing the first 10 eigenvalues of corrected adjacency matrix	1 (1); 2 (1)
2D-autocorrelations (2DAUTO)	Molecular descsriptors calculated from the molecular graphs by summing the products of atom weights of the terminal atoms of all the paths of the considered path length (the lag)	1 (1); 2 (26)

* Models emerged from CP-MLR protocolwith filter-1 as 0.3; filter-2 as 2.0; filter-3 as 0.74; filter-4 as 0.3 ≤ Q² ≤ 1.0, number of compounds in each dataset was 47.

Table III.  Descriptors identified for modeling the malonyl coenzyme A decarboxylase inhibitory activity of N-alkyl-N-(1,1,1,3,3,3-hexafluoro-2-hydroxypropylphenyl)amide derivatives along with the average regression coefficients, standard deviation and the total incidence.

Descriptor^*	Avg reg coeff (sd) total incidence^†	Descriptor^*	Avg reg coeff (sd) total incidence^†
TOPO
X2A	− 27.459 (–)1	X0Av	13.395 (–) 1
IVDE	1.959 (–) 1	BIC2	10.675 (1.247) 4
IC1	1.247 (–) 1	IC3	0.794 (–) 1
IC2	2.037 (0.221) 5	SIC3	10.696 (–) 1
TIC2	0.013 (–) 1	BIC3	10.913 (3.004) 3
SIC2	10.694 (1.654) 3	CIC5	1.826 (–) 1
CIC2	− 1.462 (0.535) 3
BCUT
BELm3	3.327 (0.224) 2	BELm6	3.406 (–) 1
GVZ
GGI2	1.246 (–) 1	JGI8	152.478 (9.716) 4
GGI4	1.240 (–) 1	JGI9	122.281 (–) 1
GGI8	3.453 (0.083) 2
2DAUTO
ATS5p	0.045 (0.011) 2	MATS8e	5.769 (0.505) 4
MATS4m	− 61.449 (2.748) 2	GATS4e	3.261 (–) 1
MATS5m	− 47.798 (–) 1	GATS6e	− 3.611 (0.227) 8
MATS7m	21.965 (5.034) 4	GATS7e	− 2.256 (0.729) 3
MATS6e	6.475 (–) 1	GATS8e	− 2.466 (0.174) 2
MATS7e	2.599 (0.713) 2

* The descriptors are identified from the one and two parameter models emerged from CP-MLR protocolwith filter-1 as 0.3; filter-2 as 2.0; filter-3 as 0.76; filter-4 as 0.3 ≤ Q² ≤ 1.0; number of compounds in the study was 47; X2A and X0Av are the average connectivity indices (chi-2 and chi-0, respectively); IVDE is the mean information content on the vertex degree equality; BIC2 and BIC3 are the bond information contents (neighborhood symmetry of 2- and 3-order, respectively); IC1, IC2 and IC3 are the information content indices (neighborhood symmetry of 1-, 2- and 3-order, respectively); SIC2 and SIC3 are the structural information contents (neighborhood symmetry of 2- and 3-order, respectively); TIC2 is the total information content index (neighborhood symmetry of 2-order); CIC2 and CIC5 are the complementary information contents (neighborhood symmetry of 2- and 5-order, respectively); BELm3 and BELm6 are the lowest eigenvalue n = 3 and 6 of Burden matrices, respectively / weighted by atomic masses; GGI2, GGI4 and GGI8 are the topological charge indices of order 2, 4 and 8, respectively; JGI8, JGI9 are the mean topological charge indices of order 8 and 9, respectively; ATS5p is the Broto-Moreau autocorrelation of a topological structure–lag 5/weighted by atomic polariazabilities; MATS4m, MATS5m, MATS7m are the Moran autocorrelations-lag 4, 5 and 7, respectively / weighted by atomic masses; MATS6e, MATS7e, MATS8e are the Moran autocorrelations-lag 6, 7 and 8, respectively / weighted by atomic Sanderson electronegativities; GATS4e, GATS6e, GATS7e, GATS8e are the Geary autocorrelations-lag 4, 6, 7 and 8, respectively / weighted by atomic Sanderson electronegativities; also see ref. (7) † the average regression coefficient of the descriptor corresponding to all models and the total number of its indices; the arithmetic sign of the coefficient represents the actual sign of the regression coefficient in the models.

Table IV.  Predicted residual activity of three test-sets (14 compounds in each) derived from the compounds in Table I.

	Residual^*
Compound number	Des Clus^†	Act Clus^‡	Rand Clus^¶
1			0.25
2	0.00	− 0.20
3	− 0.16
4			− 0.05
6		− 0.23
7			0.09
8	0.49
11	0.10	0.28	0.12
12^§	5.55	5.73	5.53
15			0.06
16		0.77
18	0.13
19	7.31	7.22	7.32
20			− 0.15
22		− 0.77
23			− 0.08
24		0.30
26		− 0.09
27			0.27
28	− 0.16	− 0.21
29^§	4.95	5.11	4.83
30			0.09
31^§	6.04	6.28	5.96
32	0.67	0.79
34	− 0.30
36		− 0.07	0.02
38	0.36	0.33
40			− 0.01
41	− 0.47
42	0.23	0.30
44			0.24
45	0.00
46		− 0.46	− 0.65
47	− 0.06
48	− 0.08
50		0.25	0.31
Test set R^2	0.841	0.678	0.850

* The difference between observed and predicted –logIC₅₀ values; the training models have 32 compounds each ^† test-set from the cluster analysis of all descriptors used in deriving (Equation 3) ^‡ test-set from the cluster analysis of the activity of the compounds ^¶ test-set from random selection of the compounds ^§ predicted activity for the compounds which are reported inactive at 10 μM.

Figure 1.  Plot of observed versus calculated –logIC₅₀ values.

Figure 2.  Plot of observed –logIC₅₀ versus residual values of training and test sets developed from (A) Descriptors, (B) Activity values and (C) Random selection of the compounds.