The Moriguchi MLogP is calculated from a model consisting of a multiple regression equation based on 13 structural parameters [I.Moriguchi, S.Hirono, Q.Liu, I.Nakagome, and Y.Matsushita, Chem.Pharm.Bull. 1992, 40, 127-130; I.Moriguchi, S.Hirono, I.Nakagome, H.Hirano, Chem.Pharm.Bull. 1994, 42, 976-978].
The regression coefficients have been evaluated by a training set of 1230 organic molecules, including general aliphatic, aromatic, and heterocyclic compounds, containing the following atoms: C, H, N, O, S, P, F, Cl, Br, I.
The statistical parameters of the model are: r = 0.952; s = 0.422; F0(13, 1216) = 900.4
MlogP = -1.041 + 1.244(CX)0.6 - 1.017(NO)0.9 + 0.406(PRX) - 0.145(UB)0.8 + 0.511(HB) + 0.268(POL) - 2.215(AMP) + 0.912(ALK) -0.392(RNG) -3.684(QN) + 0.474(NO2) + 1.582(NCS) + 0.773(BLM)
The model variables are frequencies (denoted by N) or presence/absence (denoted by D) of some molecular features. Their description is reported in the table below.
Parameter |
Type |
Description |
CX |
N |
Summation of weighted numbers of carbon and halogen atoms; the weights are: 0.5 for F, 1.0 for C and Cl, 1.5 for Br, and 2.0 for I. |
NO |
N |
Total number of Ns and Os. |
PRX |
N |
Proximity effect of N/O: 2 for X-Y and 1 for X-A-Y (X, Y: N and/or O; A: C, S, or P; -: saturated or unsaturated bond) with a correction (-1) for -CON< and -SO2N< |
UB |
N |
Number of unsaturated bonds including semi-polar bonds such as N-oxides and sulfoxides, except those in NO2. |
HB |
D |
Dummy variable for the presence of intramolecular hydrogen bond as ortho-OH and -CO-R, -OH and -NH2, -NH2 and -COOH, or 8-OH/NH2 in quinolines, 5 or 8-OH/NH2 in quinoxalines, etc.
|
POL |
N |
Number of aromatic polar substituents (aromatic substituents excluding Ar-C(X)(Y)- and Ar-C(X)=C; X, Y: C and/or H). Upper limit = 4. |
AMP |
N |
Amphoteric property; a-aminoacid = 1, aminobenzoic acid = 0.5, pyridinecarboxylic acid = 0.5. |
ALK |
D |
Dummy variable for alkane, alkene, cycloalkane, cycloalkene (hydrocarbons with 0 or 1 double bond) or hydrocarbon chain with at least 7 carbon atoms. |
RNG |
D |
Dummy variable for the presence of ring structures except benzene and its condensed rings (aromatic, heteroaromatic, and hydrocarbon rings).
|
QN |
N |
Quaternary nitrogen >N+<: 1; N-oxide: 0.5. |
NO2 |
N |
Number of nitro groups. |
NCS |
N |
Isothiocyanate (-N=C=S): 1.0; thiocyanate (-S-C#N): 0.5. |
BLM |
D |
Dummy variable for the presence of b-lactam. |
The MlogP model implemented in DRAGON has been evaluated on a set of 3576 compounds with known experimental logP taken from the NCI Open DataBase. The resulted correlation coefficient r was 0.935. Moreover, on our internal logP data set constituted by 10068 compounds the correlation coefficient r between experimental and calculated logP was 0.898.