JPH0216448A - Estimation processing method for distribution coefficient of chemical material - Google Patents

Abstract

PURPOSE:To predict the distribution coefft. of a material having a ring which does not contain satd. carbon by giving at least 5 atomic group constants to a heteroatomic group constituting the ring and an atomic group bonded to the ring and combining these constants. CONSTITUTION:The contribution component pi' by the partial structural part consisting of the atoms with which the distance from the substitution position among the change rates pi of the logarithmic value of the is distribution coefft. when one atomic group is substd. with the arom. ring is in n (1<=n<=10) pieces range in the number of bonds, the ratio alpha at which one atomic group changes the pi' of the other atomic group, the electronic substitution constant sigma of one atomic group, the scale rho at which the pi value of one atomic group changes according to the sigma value of the other atomic group, and the change rate F of the logarithmic value of the distribution coefft. when >=2 atomic groups are substd. in proximity to each other are given and the distribution coefft. P of the objective material is calculated by the equation 1. The distribution coefft. of polysubstd. benzene, etc., is predicted with good accuracy in this way.

Description

DETAILED DESCRIPTION OF THE INVENTION [Field of the Invention] The present invention relates to a processing method for estimating a partition coefficient of a chemical substance, and more particularly to a processing method for predicting a partition coefficient from the chemical structure of a substance.

[Technical Background of the Invention] In recent years, the use of chemical substances has been increasing not only in the field of organic analytical chemistry, but also in the fields of biochemistry such as enzyme reactions, environmental chemistry such as environmental transport, and medical and pharmaceutical sciences such as drugs, toxicants, and anesthesiology. The distribution coefficient is becoming very important. For example, the partition coefficient is thought to correspond to the degree to which a drug is transported within a living body and the degree to which a drug hydrophobically binds to target cells. However, it is not always easy to measure the partition coefficient of a chemical substance, and furthermore, it is necessary to actually synthesize the substance for measurement. If we were to test a compound in which substituents of the same type were introduced into naphthol, we would have to
(1) It is virtually impossible to synthesize as many compounds as possible.

In the past, when trying to design a chemical substance suitable for a specific use, the process of first synthesizing the substance and then testing it was repeated over and over again. Such trial-and-error methods had a low success rate, requiring the synthesis of thousands of chemical substances before finally yielding just one substance suitable for the purpose. In recent years, structure-activity relationship techniques have become popular, in which the chemical structural features of a substance are extracted using computers and the like, and various physical property values are predicted based on this. For example, Hammett (Hammett)
σ value of tact), E of tact (Tart), i-direction, solubility, molecular refraction, heat of combustion, heat of vaporization, boiling point, melting point, viscosity, surface tension, pressure in pressure, specific heat, critical temperature, critical pressure, critical volume,
Basic physical properties such as heat capacity and dissociation constant, as well as LD5G (
Semi-lethal dose: an indicator of the acute nature of a chemical substance), bioconcentration, sweetness, narcotics, hypnosis, hallucinogenicity, mutagenicity, carcinogenicity,
High-level physical property values such as herbicidal activity and octane number are being predicted.

Generally, in order to predict the distribution coefficient from the molecular structure, it is estimated by the sum of the contributions of each part (atomic group) of the molecular structure to the logarithm value (!; LogP) of the distribution coefficient, but this alone is not enough to predict the distribution coefficient between the atomic groups. Since interactions and interactions with solvents are ignored, large errors occur. Therefore, it is necessary to introduce various correction terms. Also,
Another problem is how to estimate the contribution of each part.

Specifically, the following documents are known regarding methods for predicting partition coefficients: [1] Hanshiyu, Leo, "Substituent Constants for Correlation Analysis in Chemistry and Biology" (published by Wiley Interscience) [ G.Hansch and A.L.
eo,”5ubstituent Con5tants
for Correlation Analysis
in Chemistry and Bio
log7, Wiley-Interscience
ce, New York, 1979].

[2] Rekker, “Hydrophobic Fragment Constant” (published by Elsevier) [R, F. Rekker, “The
Hydrophobic Fragmental
GonstantIl, Pharmacochem
istry Librar! , Vol. 1. Els
evier, New York, 1977] [3]
Hobfinger, Heterschell; Journal on Medicinal Society Chemistry [A, J.

)1opfinger and R,D,Batter
shell, J. Med, Chem.

19, 569-573 (197G)].

[4] Kramer; Journal on American Chemical Society [R, D, Craley III]
.

J,Am,Ghe+s,Soc,,102. 1837
-1849 (1980); 1 bid.

102, 1849-1859 (1980)], [5] Lewis, Mielys, Tiller;
, J. Lewis, M. S., MirrleSsand.
P, J, Ta71or, Quant, 5truct
-Act, Relat, 2゜1-6 (1983);
1bid, 2.100-111 (1983)] [
6] Kiel, Hall, "Degree of Molecular Bonding in Chemistry and Drug Research" (Academic Press) [L, B, Ki
er and L, H, Hall, “Mo1ecul
ar Connectivit7 in Chemis
try and drug research'', A
cade+sic PreSs (1976)], [7]
Dunn, Johansson, Wald
7pus [Wl, Dunn III, E, Johans
onand S, Wold, Quant, 5truc
t-Act, Relat, 2, 156-163 (19
83)], [8] ';' Roto, Morrow, Hentaiku; European Journal on Medicinal Chemistry
-[P, Broto, G, Moreau and
C, Vand7cke, Eurl, Med, ch.
e+s,, 19.7l-78 (1984) ], [9
] Takayama, Akamatsu, Fujita; Quantitative Structure-Activity Relationships [C, Tak
ayama, M., Akamatsu and T.F.
ujita, Quant, 5truct-Act, R
e1at, 4.149-160 (1985)].

[lO] Klopman, Namboodi, Chochet; Journal on Contestational φ Chemistry [G, KIopman, K, Namboodi
ri and M, 5chachet, J, C:am
put, chem, 6.28-38 (1985)] [11] Klopman, Ilov: Journal on Computational Chemistry [G, KIo
p+san and L, mouth, Iroff, J
, Compute, C:hem,, 2°157-16
0 (1980)].

[12] Kamre, Apraham, Do/\T, Taft: Journal e-On Pharmaceuticals.
Mistry - [M, J, Kaml
et, M, H, ^brahas.

R, M, Doherty and R, W, Taft;
J, Pharm, Sci, 74°807-814
(1985)] [13] Cambell, Lachy: Environmental Science and Technology [J, R.

Campbell and R.G., Luthy, E.
nviron, Sci, Tecb,.

19, 980-985 (1985)] [14] Gose, Crippen; Journal on Computational Chemistry [A.

K, Ghose and G, Grippen,
J, Gomput, Cbem, 7゜565-577
(1986)] [15] Ghose, Brichett, Crippen; Journal on Contestational Chemistry - [A.
t and G, M.

Grippen, J., Comput, Ghem.
8O-90(+988)] [16] Suzuki,
Kudo, [Collection of Abstracts of the 1st Information Chemistry Symposium J p, 1
70-173 (1987) [17] Kasai, Hosohi, “First
2nd Structure-Activity Relationship Symposium Abstracts JP, 26
7-270 (19B4) [18] Kasai, Hosohi, Umeyama, “
13th Structure-Activity Relationship Symposium Abstracts jp,
262-265 [19] Sasaki, Kubodera, Matsuzaki, Umeyama,
[Collection of Abstracts of the 15th Structure-Activity Relationship Symposium JP
.

[20] Moriguchi, Kanda, Komatsu; Chemical Pharmaceutical Φ Chemistry [1, Noguchi.

Y, Kanda and K, Kamatsu,
Chem, Pharm, Bull, 24°179
9 (1976)] [21] Kitano, Tomita, Taka B, “Chemical Products J 6.15 [2
2] Hirono, Komatsu, Omoteguchi, "Collection of Abstracts of the 10th Structure-Activity Relationship Symposium J (1983) [23] "Handbook of Chemical Properties Estimation Methods" by Riemann, Leal, Rosenblatt (published by McGraw-Hill) [W. J.Ly
+wan, W, F, Reeh! and mouth, H, Ro
senblatt"Handbook of Che
Mical Property Estimation
Method'' McGraw-Hill, N.
ew York (+982)] [24 Kocho, Jars: Journal on eChemical Information and Computer Science [J, T, Gh
ou and P, C, Jurs.

J,Cbem,IniCamput,Sci,, 1
9. 172-178 (1979)].

However, it has not yet been possible to accurately predict the distribution coefficient of polysubstituted aromatic rings such as polysubstituted benzene.

In addition, for disubstituted benzene, a highly accurate prediction method with Hammett type correction was developed by Fujita, Progress Oin Physical Laboratory, Organic Φ Chemistry - [T, Fu
jita, Prog, Ph1s, O
rg, Chew,, 14°75-113(
March 198 and Leo, Journal on Chemical
Society Parkin Transiti, N [A, Le
o, J., Cbeys, Sac,, Perkin.
Trans, II.

825-838 (described in March 1988, but it has not reached the point where accurate predictions can be made for samples with three layers or more. The latter paper by Leo also includes a sub-measurement example of the partition coefficient for benzene derivatives in samples with three layers or more. However, with this method, the calculated values may have a significant difference from the actual measured values, making it impossible to predict with high accuracy.

[Summary of the Invention] An object of the present invention is to provide an estimation processing method for accurately predicting the distribution coefficient of a chemical substance.

Another object of the present invention is to provide an estimation processing method for predicting the distribution coefficient of a chemical substance with high precision and in a short time by computer processing.

That is, the present invention is a method for estimating the distribution coefficient of a chemical substance having a ring that does not contain saturated carbon; Atomic group constant: When an aromatic ring is substituted with one atomic group, the change in the logarithm of the distribution coefficient π, the distance from the substitution position is the number of bonds (n is the number of bonds (where n is l≦n≦ contribution by a substructure consisting of atoms in the range of 10 (an integer that is lO) π; a measure by which one atomic group changes the π' value of another atomic group α; an electronic substituent constant σ of one atomic group.

The scale ρ by which the π value of one atomic group changes according to the σ value of another atomic group, and the amount of change F in the logarithm of the distribution coefficient when two or more atomic groups are substituted close to each other. , and combine these group constants to form equation (1): (where P is the distribution coefficient of the chemical substance, and !logP additive is the distribution coefficient assigned to each atomic group constituting the substance. It is the total sum of the contribution of the logarithm value and the sum of correction factors based on bonds, branches, and proximity of polar groups in the structure of each atomic group, and is
each represents an atomic group contained in the substance), and provides an estimation processing method comprising calculating the distribution coefficient of the chemical substance using the equation (1).

In addition, in the present invention, the partition coefficient P of a chemical substance refers to water (
It refers to the proportion of chemical substances dissolved in two phases: a water phase) and a liquid (oil phase) that is substantially insoluble in water, and is defined by the following formula.

Concentration of a chemical substance in the aqueous phase A cause of estimation errors in predicting the partition coefficient of a chemical substance is that substitution of strong electron-withdrawing groups such as NO2 and CN causes changes in solute/solvent interactions due to hydrogen bonding. (electron-withdrawing effect), the introduction of a substituent changes the orientation of solvent molecules, which causes the cluster structure of the surrounding solvent molecules to fluctuate (fluctuation of the solvent cluster structure), and the intramolecular hydrogen bonding of the substance itself causes the interaction between solvent molecules. When polar groups are in close proximity, the direct hydrogen bonds with solvent molecules are reduced because they incorporate water molecules from the solvent and form a stable hydrate structure (one-to-one water bonding). formation)
, resonance inhibition (d
(e-coupling of resonance) occurs (ortho effect), and a polar group such as a nitrogen atom is shielded by a nearby substituent such as a phenyl group, making it impossible for solvent molecules to approach and effectively exerting its polarity. (blocking of polar groups), because hydrophobic bonds occur in long hydrocarbon parts within or between molecules, the iceberg structure around the hydrocarbon is removed and the contact area with water molecules is reduced; changes in solute/solvent interactions due to changes in the charge distribution of the ammonium cation depending on the length of the surrounding hydrocarbon moiety in the case of ammonium salts or due to changes in the solvent cluster structure;
etc. can be mentioned.

In addition, intramolecular hydrogen bonding, one-to-monohydrate formation, polar group
Hiding etc. are okay) 2? These effects are often referred to as the ortho effect, as they are often caused by the exchange of

Conventionally, when estimating the partition coefficient of an aromatic ring compound having a substituent, the partition coefficient is divided into a term based on the individual substructures constituting the compound and a correction term based on the interaction between each substructure. A method of estimating by summation has been proposed. For example, dibenzene (X-Cb Ha-Y)
In Leo's paper, the partition coefficient of is calculated based on the individual substructures as follows:
and the sum of the amount of change Δτ×v in the π value of the substituent. Δπ8ν consists of a correction term ρVσ8 based on Hammett's σ value and a correction term F based on the ortho effect.

1 ogP = l ogP cbo6+ tr ×+
πy+Δπxy1 ogP addilve= l
ogP C6H6 + π × + πyρyσ × 10F = Δ5V (where P and P C6H6 are the partition coefficients of disubstituted benzene and benzene, respectively, and π × and π,
are the π values of the substituents x and Y, respectively, and πX = l ogP C685X - sentence ogP
C6H5 and πV: logP C6H3Y -1ogP
cbHb) The method for unreleased IJ1 is as follows: a) In addition to the correction term described in J-, a correction term is introduced that takes into account that the π value of one substituent changes depending on other substituents, and b) ) Furthermore, the ρ determined for each atomic group can be changed depending on other nearby atomic groups, taking into consideration the environment in which that atomic group is located.

When introducing the correction term in a) above, that is, a correction term that takes into account that the π value changes depending on other substituents, rather than targeting the π value of the entire substituent, It is preferable to consider only the affected part because of its accuracy. Therefore, in the present invention, among the π values, the distance from the substitution position is n in the number of bonds (where n is an integer satisfying l≦n≦lO). ) is introduced into the correction term.

According to the method of the present invention, the above ((
By estimating based on the formula 1), it is possible to accurately predict the distribution coefficient of a polysubstituted product such as an aromatic ring or a heterocycle having two or more substituents, particularly three or more substituents. That is, the sentence ogP addi which is a term based on each substructure
tive (in the fragment method, this is obtained as the sum of the contribution of each atomic group and the correction term based on the structure of each atomic group), as well as the Hammett's σ value etc. as a correction by multiple substitutions. By adding the amount of change due to the amount of change based on the π' value of the group and configuring the parts other than logP additive, various effects that affect the distribution coefficient, such as the electron-withdrawing effect, fluctuation of the solvent cluster structure, ortho effect, etc. The distribution coefficient can be predicted by taking into account the effects of As a result, it is possible to obtain a predicted value of the distribution coefficient within the range of the actual measured value and a tolerance (generally ±0.2).

In addition, the atomic group constants α ρl, σJ, and F i in equation (1) are determined by performing statistical processing such as multiple regression analysis based on a large number of actually measured values for each atomic group. By using these values, the distribution coefficient of a chemical substance can be predicted with even higher accuracy.

Furthermore, by representing the chemical structure of a substance in the form of a bond table that includes information about atoms and the bonds between these atoms, it becomes possible for computers to recognize aromatic rings, substituents, etc. This can be done automatically through processing. As a result, no matter how complex the structure of a chemical substance is, or even if it is an unknown structure that has not been synthesized or discovered, the distribution coefficient can be predicted with high accuracy in a short time.

According to the present invention, it is possible to shorten the time required for information collection, investigation, etc. in individual research by engineers, and increase the density of information obtained, making it possible to proceed with research efficiently. . The obtained information on partition coefficients can be advantageously applied to structural analysis of chemical substances, molecular design, organic synthesis route design, etc., which are highly requested by engineers involved in drug manufacturing. .

Preferred embodiments of the present invention are listed below.

(1) The above atomic group constants α, ρ, σ and F are respectively
A method for estimating a distribution coefficient of a chemical substance, characterized in that the optimum value is calculated by performing statistical processing for each atomic group.

(2) A method for estimating a distribution coefficient of a chemical substance, wherein the atomic group constant σ is a Hammett's σ value.

(3) The atomic group constant π is a value obtained as the difference between the logarithm of the distribution coefficient of an aromatic ring substituted only with the atomic group and the logarithm of the distribution coefficient of an unsubstituted aromatic ring. A method for estimating the distribution coefficient of a chemical substance.

(4) The above nogP addiLive calculates the sum of the partial quartiles of the logarithm values of the partition coefficients assigned to the substructures divided by isolated carbons (carbons that do not have unsaturated bonds with heteroatoms) and each part. A method for estimating a distribution coefficient of a chemical substance, characterized in that it is a total sum of correction factors based on bonds, branches, and proximity of polar groups in the structure.

(5) The above 1 ogp addiLive calculates the logarithm of the distribution coefficient of the unsubstituted ring containing no saturated carbon contained in the chemical substance and each heteroatom group constituting the chain ring and/or each bonded to the chain ring. the π value assigned to the atomic group,
Proximity of bonds, branches, and polar groups in the structure of each atomic group,
A method for estimating a distribution coefficient of a chemical substance, characterized in that it is a sum of a total sum and a sum of correction factors based on .

(6) The chemical substance having a ring that does not contain saturated carbon is
A method for estimating the distribution coefficient of a chemical substance characterized by being an aromatic ring compound, an aromatic heterocyclic compound, or a condensate thereof.

(7) The chemical substance having a ring that does not contain saturated carbon,
A method for estimating the distribution coefficient of a chemical substance, characterized in that it contains two or more heteroatomic groups constituting a ring and/or two or more atomic groups bonded to the ring.

(8) The chemical substance is given as a bond table containing information about the atoms constituting the substance and the bonds between these atoms, and based on this bond table, a ring that does not contain saturated carbon,
Detect the heteroatomic group constituting the chain ring and/or the atomic group bonded to the chain ring and, if necessary, the partial structure of the substance, and then use the atomic group constant corresponding to the detected atomic group to detect the atomic group according to formula (1). A method for estimating the distribution coefficient of a chemical substance, the method comprising calculating the distribution coefficient of the substance.

(9) The above π' value is based on the sum of the contributions of the logarithm of the distribution coefficient assigned to the substructures divided for each isolated carbon, and the proximity of bonds, branches, and polar groups in each substructure. A method for estimating a distribution coefficient of a chemical substance, characterized in that it is a total sum with a sum of correction factors.

(lO) A chemistry characterized in that the contribution of the logarithmic value of the partition coefficient assigned to the partial structure divided for each isolated carbon is multiplied according to the type of isolated carbon to which the partial structure is directly bonded. A method for estimating the partition coefficient of substances.

[Structure of the Invention] In the method for estimating the distribution coefficient of a chemical substance of the present invention,
The partition coefficient P represents the distribution ratio of a chemical substance dissolved in two phases consisting of an aqueous phase and an oil phase, and is expressed by the following formula: 0 defined by the concentration of a chemical substance in the aqueous phase. Logarithmic value (dan og
P) is displayed.

As the oil phase liquid, n-old ethanol, benzene, chloroform, hexane, cyclohexane, etc. are used, and n-octanol is particularly common. In the aqueous phase,
In order to keep the ionic strength constant, electrolytes may be added or P
A buffer may be added to keep H constant. Unless otherwise specified herein, the partition coefficient is n-
This is the value for the ethanol/water system. Also, the concentration of a chemical substance is usually expressed in moles/su.

In the present invention, the estimation formula used to calculate the distribution coefficient is basically expressed by formula (1).

(However, lagP: logarithm of the distribution coefficient of a chemical substance ogP ad
ditive: The sum of the flagP′a portions assigned to each atomic group constituting the substance and the structure of each atomic group. The total sum of the correction factors based on one branch of the bond and the proximity of polar groups α1: The measure by which atomic group i changes the π° value of atomic group j π, ′: The exchange of atomic group i into an aromatic ring The contribution of the substructure part consisting of atoms whose distance from the substitution position is within the range of n bonds (where n is an integer satisfying 1≦n≦10) to the change amount π of the intersection ogP when ρ1: A measure by which the π value of atomic group i changes according to the σ value of atomic group j σJ: Electronic substituent constant F of atomic group i i, j, :rX subgroups i, j,... The amount of change in ogP when closely substituted (i, j, . . .: atomic groups contained in the substance) can be derived by a procedure such as a correction term, which is a characteristic feature of the present invention.

It has been found that for a series of compounds having similar hydrogen bonding moieties, the association constant PKHB due to hydrogen bonding and the acid dissociation constant pKa exhibit a good first-order correlation. Further, according to Hammett's law, pKa can be correlated with Hammett's σ value using a linear equation.

The π value of the substituent (fLogPx −uogPu; where 1 ogP H is the standing og of the compound in which the substituent
Similarly, it has been revealed that there is a first-order correlation with the Hammett's σ value for the P value (P value), and the following formula generally holds; however, the R substituent Y is a non-hydrogen bonding substituent. .

Δπ=πX/PhY−πX/PhH” CI σx+C
2 where πX/PhY and πX/Ph)l are respectively substituted, 7;! π value for Y-substituted benzene of ix and π value for substituted 24i X benzene: πx/phv = l ogP XphY −1agP
HphYπx/ph+ = n ogP XphH−
l ogP C6H6 (however, P to phB is A-C6
represents the distribution coefficient of Ha-B), σX is the σ value of substituent X, and C1 and C2
are constants.

From the above, it is thought that the change in π value depends on the hydrogen bond between the solute and the solvent. Based on this idea,
For disubstituted benzene (X-C6H4-Y), Fujita calculated formula (2): %formula% (where σY is the σ value of substitution 2!iy, aY, ρ
X and ρY are each constant) is proposed. aY has a value slightly smaller than l, which is thought to be due to a change in the cluster structure of water molecules around the solute molecule.

Undeveloped Ell # is disubstituted/(7ze7(X-C6H4-Y
), we first expanded equation (2) for imxy as well.

πY/PhX ” a Xπ1PhH+ρXσY×ρY
σX... (3) (However, ax is a constant) The lagP value of this compound is derived from equations (2) and (3) as follows, respectively.

l ogP xphy ; sentence ogP C6N6+πX/PhY+πY/PhH see ogP C6H6+πX/PhH+πY/PhH+
(a y-1)w xlphH+ p xσY+ρYσ
Xfl ogP yphx = l ogP C6H6+π1PhX + ff
X/PhH= l ogP C6H6+πY/PhH+
πX/Ph)I+ (a x-1)wylphH+
p
'//PhH+ρ× σY +ρY σx
(6) Here logP additive
= l ogP cbsb+πX/PhH+tcv/p
hs, αX = (ax-1)/2, a
If we set v = (ay-1)/2, 1 ogP xphy = l ogP additive +αYπX/PhH+αXπY/Ph)l+ρ× σY
+ρY σx (7) Here, l o
gP add ive is the sentence ogP value calculated assuming additivity. αX is a measure of how substituent X causes a change in the π value of substituent Y, and is considered to correspond to a change in the cluster structure of solvent molecules. Here, if the substituents X and Y are large, rather than correcting the original π value by multiplying it by α,
Select a collection of atoms whose distance from the substitution position is within the range of n bonds (n is an integer where l≦n≦10) as a part that is subjected to the influence of α, and calculate the π value of that part. contribution π
It is more accurate and preferable to correct the value by multiplying it by α.
Therefore, πX/PhH and ff Y/PhH in equation (7) are
Let ff' X/PhH and π' 1PhH. A value of 3 to 6 is particularly suitable for n. 22If the substituent is small, π
and π° will have the same value, but will have different values for large substituents.

It has been suggested that the additivity of Hammett's σ value holds for the acid dissociation constant and the relative reaction rate constant. (7
) Assuming that additivity also holds for α in the formula, formula (8) is derived for polysubstituted benzene.

1ogP=logP C686+Σπ1= l ogP
additive (where i and j indicate atomic groups contained in the target substance) Furthermore, in equation (8), a correction term ΣF i due to the ortho effect,
By adding r and -, an equation corresponding to the above equation (1) is obtained. Here, when substituent i is close to substituent j, if ρ of substituent i can be changed depending on substituent j, unique mutual interactions such as intramolecular hydrogen bonds between substituents Equation (1) holds true even in poems where action exists, and the obtained equation can be applied not only to polysubstituted benzenes but also to polysubstituted rings that do not contain saturated carbon, such as aromatic rings and heterocycles.

Therefore, the chemical substances to which the method of the present invention is applied are organic compounds having rings that do not contain saturated carbons. Examples of rings that do not contain saturated carbon include aromatic rings such as benzene and biphenyl; fused aromatic rings such as naphthalene and anthracene;
Examples include aromatic heterocycles such as pyrazole and pyridine; and fused heterocycles with quinoline, quinoxaline, and phenanthroline.

It is preferable that the chemical substance contains two or more atomic groups, including a hetero atomic group constituting a ring that does not contain saturated carbon and an atomic group (substituent) bonded to a chain ring. The method of the present invention can be suitably applied when three or more atomic groups are included. Examples of heteroatomic groups are -NH-0-, -S-, -5e--N=N--NH-
N-N-NH- (H=N-, -9O2N-OH-NH-
etc. can be mentioned. In addition, atomic groups bonded to rings include those consisting of a single atom, such as halogen 7. Examples include OH, NH2, CN, an alkyl group, an aryl group, a carboxylic acid group, an amide group, and the atomic groups may form a ring.

The atomic group can be arbitrarily selected for each chemical substance, but the atomic group cut out by cutting the bond bonded to the isolated carbon (carbon that is not unsaturated with a heteroatom) IX child according to the fragment method is used. This is preferable because it is convenient for processing.

When there are multiple limit structures, such as a pyridine ring, if at least one of the limit structures has a carbon atom that is not unsaturated with a heteroatom, that is also considered an isolated carbon. A plurality of atomic groups may be combined to form a new atomic group.Furthermore, a single bond originating from a carbon within an aromatic ring may also be a target for cleavage, if necessary.

Examples of atomic groups selected for chemical substances are shown below (the part surrounded by +' and -: atomic groups).

Below Margin Below Margin Next, various constants including the atomic group constant appearing in formula (1) will be specifically explained.

(A) Method for determining π π' In equation (1), the atomic group constant π1' is the result of the substitution of atomic group i into an aromatic ring composed of hydrocarbons
Of the change in P value π, this is the contribution from the substructure consisting of atoms whose distance from the substitution position is within the range of n bonds (where n is an integer satisfying 1≦n≦10). . Atom group I also includes a heteroatom group within a heterocycle. The π' value can be determined in the same manner as the known π value determining method described in the above-mentioned literature.

As an example, the π' value of the atomic group OH for benzene can be obtained by subtracting the lagP value of benzene from the l age value of phenol.

π OH= sentence age()! noh) -also ogP
(Benzene) Similarly, regarding the heteroatom group constituting the ring such as -N= in pyridine,
Difference between gP value and benzene intersection with gP4: π'-N・
Two ogP (pyridine) - JlagP (benzene)
obtained as. These two examples show that the atomic group OH and the atomic group -
Since N= is also a small atomic group, the π value and the π' value are the same.

Alternatively, the π' value can also be obtained by calculation based on the fragment method described below.

If the substituent is small and the π and π' values are equal, then the π' value can be set to, for example, Hansch, Leo, Unger, Kim, Nikaitani, and Lee; ,
H, Unger, H, Kim, D, N1kaita
ni and El, Lien, J., Wed.
Chew,, 16. 1207 (1973)
] The π value described in can be substituted.

Determination of π' by the fragment method The π' value can be determined by the fragment method similarly to the π value. According to the fragment method, it is possible to obtain the π' value by combining i1 of each factor, which will be described later, so it is most convenient for implementing final complete IJI using a computer. The fragment method is described in detail in, for example, the above-mentioned documents [1], [2], and [23]. As described in these documents, when calculating the π゛ value by the fragment method, we do not simply sum up the (logP ′B fraction of each atomic group, but also calculate 〈 Correction factors need to be included.

According to literature [13, [23], the π° value is calculated from the following factors.

fx: Sentence o of atomic group X (including those consisting of two atoms)
gP contribution.

fdX: Log when atomic group X is substituted at one place on the aromatic ring
P contribution (however, fllM indicates that the atomic group X is contained in the aromatic ring) fdll×=the sentence ogP contribution when the atomic group X is substituted in the aromatic ring at two places.

Fb: Correction of bonds other than aromatic rings.

Fe6r: Correction of branches consisting only of carbon.

FgB": Correction of branches containing hydrogen-bonding groups.

Fpn: Correction for hydrogen-bonding groups being close to each other via n isolated carbons.

F　bYNC tertiary amine bond correction.

FbYPn: Correction of phosphate ester bond (n indicates that n atoms are connected to the substituent position. References [1] and [23] make corrections unrelated to n, but depending on n The accuracy of the estimation will be improved if the correction is made using

F x M) IG: E (correction for the proximity of the group X to the halogen.

Although the π° value is calculated from each factor as described above, it is preferable to consider corrections a) to f) listed below in addition to these factors.

a) Correction due to proximity of polar group and unsaturated bond: This will be described later in Example A-4 below.

b) Correction for endocyclic benzyl carbons: Depending on the number of endocyclic benzyl carbons, add, for example, 0.195 per benzyl carbon.

C) Correction due to proximity of polar group and aromatic ring: Example B- below
This will be described later in Section 3.

d) Further correction for tertiary amines.

: Described later in Example B-4 below.

e) Correction for hydrophobic binding.

Consideration is given to the effect that the easily mobile hydrophobic part within one molecule assumes a conformation that minimizes the contact area with water. For this reason, a new parameter is introduced, but since the accuracy of this correction is low, it is used as a warning against accuracy.

f) Exception handling regarding proximity correction term: (t) -X-(CH2) n-Y (X,
For compounds of the type (Y is a polar atomic group), the proximity correction described in 1 literature [1], [23] causes a large error, so a proximity correction constant is determined individually.

The fLogP distribution of the above-mentioned atomic groups, that is, the various f
The actual value of must be determined by taking into consideration what kind of isolated carbon each atomic group is bonded to. For this reason, isolated carbons are classified into the following types, and depending on which type of isolated carbon each atomic group is bonded to,
Determining f in advance improves accuracy.

1: alkyl carbon, 2: aromatic carbon, 3 nistyryl carbon, 4: vinyl carbon, 5: benzyl carbon, 6: w1 ligated carbon (may be substituted with 2), for example, fO
For H, 1-OH, 2-OH, 3-0H14-O
H, 5-0) -1, 64, -0 for each of 1
, 44-0,84, -1,24, -1,16.

The value of Fb is also changed depending on the type of bond (single bond, double bond, triple bond).

For specific values of each factor above for each atomic group, please refer to the statement #[
1], [23].

π' when n=5 considering each factor as above
An example of calculating the value is given below.

Example A knee 2 Example A-1 (However, φ represents a benzene ring) π' = fllc++ +2 fcu3
+2Fb +Fc8r -f'H (0,345)
(0,89) (-0,12) (-0,13)
(0,225)=1.530 Since π' is a part of the amount of change π in the lagP value due to the substitution of the atomic group, the contribution molecule 〆H due to -H before substitution must be subtracted.

In addition, in this example, since n=5, π = π.

below,'? ? [':1 (0,225) As in this example, when the substituent forms a ring.

It is convenient to divide in this manner according to the position from the substituent, and to equally divide the atomic groups and bonds that span both.

Example A-3 As in this example, one OH is included up to the fifth neighbor,
When the other is not included, the FPn term is multiplied by 1/2.

Example A-4 This example uses the correction term F(, ) due to the double bond and the above ra
) This is an example including the section "Correction due to proximity of polar group and unsaturated bond" (underlined part).

For the OH of ethanol and the OH of phenol, fOH is -1,64 and -0,44, respectively, and the polar group has π
When electrons come close to each other, their f value increases significantly. As mentioned above, this effect can be expected to some extent by classifying the bonding carbon into six types and determining the f value for each case.
Also consider isolated carbons adjacent to unsaturated bonds, such as =CH2 in Hx, C-CIh-OH, etc.

For example, CH2=CH-C:H2-X, φ-GH=C
H-CH2-X, CH= C-CH2-X, φ-C
In the case of having a part such as H-X (representing a φHachensenφ ring or a polar group), the correction factors are 0.252 and 0.5100, respectively.
．． The underlined part of Example A-4, which adds 698 and 0.212, is this correction factor.

(B) logP additive
Additive determination method 1 ogP additive
e is a value obtained by summing the logP contribution of each atomic group constituting the chemical substance and a correction factor based on the structure of each atomic group.

For example, it is obtained as the sum of the lagP value of the mother nucleus (unsubstituted aromatic ring) and the π value of the atomic group substituted for it. - 1 ogP addiLiv of p-nitroaniline as an example
e is 1 ogP additive= fL og
P (penzesi) + π 882 + π N02
It is obtained by

Determination of JlogPad additive by fragment method 1 The ogP additive may be calculated based on the fragment method. This calculation is performed in the same way as the calculation of π° described above. Of course, unlike the calculation of π', the The contribution molecule O1+ due to -H is not subtracted.

Example B-1 Taking p-nitroaniline as an example, first there is a carbon that does not have an unsaturated bond with a heteroatom (this is called r isolated carbon 1).
By cutting the bonds around , the following eight fragments (
substructure) into (coned part two fragments).

= 0.65 (however, f'c+: lagP contributing molecule of the atomic group CH in benzene Ilc: statement ogP of the C atom (non-bonded with H) in benzene
Contributing molecule 'N+u: atomic group NH2 substituted with benzene ogP Contributing molecule 'so2: atomic group N02 substituting o
gP contribution) Example B-2 (p-nitrophenoxyacetic acid) 1 ogP
additive (P-nidomitsuenoxy acetic acid) Then add the logP contributions of each fragment.

1 ogP additive (P-nitro7nirisi) = f'cn X 4 + flc X
2 + f'NH2+ r'No2<0.3
55) (0,13)(-1,00)(-0,03)
= 1.072 (However, f, IIo: LogP contributing molecule of atomic group 0 substituted with benzene c+z: logP contributing molecule coou of atomic group CH2
: LogP contribution of atomic group C00H Fp+: Correction for the fact that atomic group 0 and atomic group C0OH are adjacent to each other via one isolated carbon Fb: Bond between atomic group 0 and atomic group CH2, atomic group CH2 and atomic group C0OH This is a correction for. ) In this example, the correction term described in the section describing the determination of π' by the fragment method is taken into account.

Example B-3 φ-', CH2-CH2-NH2) fl ogP additive = 5 f IIcH+ f 'c+ f'
cu2+ f CH2(0,355) (0,1
3) (0,475) (0,66) + f Ns2
+ 2 F b +0.162 (-1,54)
(-0, 12) This example is the correction due to the proximity of the polar group and the aromatic ring described above in the section on determining π' by the fragment method.
This is an example that includes terms (underlined parts).

That is, for compounds containing the structure φ-(CH2) n-X (X is a polar group), when n is 2 and 3, the correction is 0.162 when n = 2 and 0.113 when n = 3. Add value. When n is 1, this is the case where X is bonded to the benzyl carbon, which can be taken into consideration by changing the value of f depending on the carbon of the above-mentioned bonding group. Further, when n is 4 or more, there is no need for correction.

In this example, there is only one substituent, so when estimating lagP, there is no need to consider the correction term due to the interaction of substituents.Therefore, the logP ad calculated here
ditive becomes the estimated lagP as is. The estimated value above is 1.42, while the actual value is 1.41.
The estimated values reproduce the actual values well. The logP add example shown below can also be directly compared with actual measured values.

= 1.42 Example B-4 1ogp additive (= logP) This example is a further correction for tertiary amines and phosphate esters described in the above section of determining π' by the fragment method. This is an example of what we did.

According to 1 literature [1], [23], the bonds in R1-N-R3 (R, , R2, and R3 are hydrocarbon fi) are corrected by (number of bonds - 1) XFbvs, but the accuracy is In order to increase the accuracy, it is recommended to limit the bonds subject to this correction to those included from the center N to the length of the shortest chain among R1, R2, and R3. In this example,
The shortest button is 1003, so 1 in the above diagram.
> was considered as the target for F bYN.

(C) σ The atomic group constant σi is an electronic substituent constant of the atomic group i, and for example, the σ value of σ (Ha) is used. Hammett's σ values include σ・, σσO1σ1, σR1
σ°, etc., and any of them can be used alone or in combination. Further, the σ value usually takes into consideration the substitution positions at the meta and para positions, and the σ-value is used for meta substitution, and the σ value is used for para substitution. For ortho substitutions, the σp value is substituted. For example, the σ value is as follows: Hansch, Leo, Ango, Kim, Nikaitani, Lee Ann: Journal on Medical Chemistry [G, Hansch, A, Leo.

S, H, Unger, K, H, Kim, D, N1k
aitani and E, J., Lien J., Med.
Ches., 16.1207 (March 197) and I1 described in "Hammett's Rule - Structure and Reactivity" by Naoki Tanemoto (published by Maruzen Co., Ltd.) can be used.

Furthermore, the unknown σI value can be determined by optimization processing such as multiple regression analysis, as in the case of the atomic group constants α1 and ρX.

(D) α and R1 The atomic group constants α1 and R1 are the scale by which atomic group i changes the π value of other atomic group j, and the π value of atomic group i depending on the σ value of other atomic group j, respectively. It is a scale that changes depending on the α
! For example, an optimized value for R1 can be obtained by performing statistical processing such as multiple regression analysis.

That is, it can be determined using the multiple regression equation: % formula % (where i, j, and k each represent an atomic group contained in the chemical substance). Here, la of the first term on the left side
By substituting the measured values of several compounds containing atomic group i into gP, substituting known atomic group constants into the second to fourth terms, and using Σπ and Σσ on the right side as explanatory variables, R1 and ρl are obtained as regression coefficients.

Computer-based iterations are used to ensure that the σ values of atomic groups whose σ values are unknown can be determined, and that α, ρ, and σ are balanced for each atomic group and consistent as a whole. You may obtain each atomic group constant by

One of the features of the present invention is that ρ is changed taking into consideration the effect of neighboring atomic groups, and this will be described later.

(E) Fi,j atomic group constant F Lj, - is two or more atomic groups i, j,
... is the amount of change in the lagP value due to the presence of ortho positions, peri positions, etc. in close proximity to each other.

- For example, if the atomic groups OH and CH3 are present in the ortho position.

F"orl, CH3= l ogP (a-methylphenol) - ogP (p-methylphenol)
obtained as. Furthermore, the optimum value may be determined by performing statistical processing such as the least squares method.

Alternatively, F i,j, may be calculated from the Swain-Lupton field constant γ and Taft (7) σI value, ES value, etc. 0 For example, a constant af for each atomic group. , given δ.

Fi, j”fi σ+j+ f j σ11+δ+Es
J+δ''Es+. (Here σI is σ”, σ.

For example, calculate as F Cool, coco3 = -0,74. However, if there are specific interactions (for example, intramolecular hydrogen bonds), it is necessary to determine them individually. For example, -OH and -COOH1 -OH and -
F on coo11 when COOCH3 is in the ortho position = 1.007, F OH,
An example is COOCH3: 0.97.

As for the σI value, Churton, Progress in Physical Organic Chemistry [M, Ghart
on, Progr, Ph7s, Org, Chem,
, 13. 119 (1981N, Bijiro, Wrecker, Reontative Straft Activity, etc.)
Relationships [G, J, Bijiloo an
d R,F, Rekker, Quant.

5truck-Act, Re1at, 3.9l-9
6 (1984), 1bid,, 3゜111-115
(The values listed in 1984 can be effectively used.

As for the ES value, Anger, Hansch, Progress in Physical Organic e-Chemistry - [
S,)1. tlnger and C,) Ians
ch, Progr, Phys.

Org. Hatsch, Journal of Medicinal Chemistry [E, Kutter, C,H
ansch, J.Med.Chem., 12.64
7 (as in September 1969), the ES value can be estimated from the 77n der Waals radius, or it can be obtained by regression analysis.

Note that F ,,j,,, is valid not only when two atomic groups are close to each other, but also when three or more atomic groups are close to each other, as shown below. It can be obtained as the sum of Fi,j for . In order to further improve accuracy, the correction value may be determined for F +, j, etc. (the part surrounded by - to 9: atomic group).

In this way, the atomic group constants α, ρ, σ, and F are determined, respectively.

The atomic group constants α, ρ, σ, F, f, and δ described above can be converged and determined by repeating the method of least squares so as to be consistent for any atomic group.

From this, πσ1. These parameters, including ES, can be summarized as a ``consistent substituent constant J''.
(Setf-consistent 5ubsitu
ent Paras+eter (hereinafter abbreviated as SSP).

Specific examples of SSP are shown in Tables 1 and 2. In Table 1, the number at the end of the atomic group indicates the type of isolated carbon to which the atomic group is directly bonded, l: alkyl carbon, 2: aromatic carbon, 3 Nistyryl carbon, 4: Vinyl carbon, 5: Benzyl carbon, 6: Condensed position carbon, 0 For example, 2-Br means φ-Br, 5-
Br means φ-CH2-Br. Also, 3-N
(1,1 is CH30H3 φ-GH=(H-N-OH3, 5-N(1,1 is φ-
Represents GH2-N-CH3.

Also, 2-OPO(0-1)2 0PO(0-1)(0-3) are each -OH3 〇－ φ means.

Also, in Table 2, Especially Kotowara Unless, The atomic group is located at the ortho position of the benzene ring. represents something to do.

Remains below 1 Table 2 Table 2 (Continued) (F) ρ Taking into account the effect of neighboring atomic groups Because certain atomic groups form intramolecular hydrogen bonds and l:l hydrates when they are in close proximity. In some cases, the ρ value changes greatly, and it is necessary to improve the prediction accuracy by changing the ρ value depending on the neighboring atomic groups.

Specific examples of the ρ value and F value in such cases are shown in Tables 3 and 4.
In Table 3, the trisubstituted atomic group is
In the table, the case of trisubstituted atomic groups is shown, however,
In Table 3, unless otherwise specified, the atomic groups represent those located at the ortho position of the benzene ring.

In addition, when atomic groups exist in different constituent rings, such as the 1st and 5th positions of a naphthalene ring, prediction accuracy can be improved by using different α and σ values than when they exist in the same ring. can be increased. For atomic groups that have been replaced with adjacent rings, it is recommended to multiply the α and σ values by 0.519.Furthermore, in the case where there is a ring intervening, such as the 1st and 6th positions of an anthracene ring, It is recommended to multiply by 0.3 for
As the σ value, σ or σ value is used based on the resonance mode of the ring.

Further, by setting an upper limit in advance to the total value of the second to fourth terms in equation (1), it is possible to prevent the total value from becoming too large. This may be set depending on the number of related atomic groups. For example, if the number of related atomic groups is 4.
, 5, and 6, the upper limit is 2.
Use 80.3.56.4.27.

In addition, consistent atomic group constants are not limited to the values shown in Tables 1 to 4, and may be modified through statistical processing, etc. in order to improve the prediction accuracy as the actual measured data of the ogP value increases. It is preferable to correct them sequentially.

Order 1' below Table 3 Table 3 (continued) Atomic groups i, j OH, (:0OH O) 1.11:0CH3 0H, (:00CH+ 0)1. C00PH OH, CHO O)1. C0NHN)12 0H,C:0NH2 0)1. NHCO(PH) Leaf, Nl[ONH(PH) OH,N02 NN2. C00H NN2. COCH3 N1-12. C00(:N3 NN2. C0N)IN82 NHz, (:00(:N2 N112. C0NH2 NN2,0H NN2. Ni12 F + J 1.007 0.970 +, 100 0.803 0.911 0.980 1.266 0.254 o, oo.

o, oo.

O,570 1,160 1,160 0,571 0,901 0,576 0,330 o, oo.

o, oo.

o, oo.

o, oo.

o, oo.

o, oo.

o, oo.

O,580 o, oo.

O, 345 o, oo.

o, oo.

o, oo.

o, oo.

o, oo.

o, oo.

o, oo.

o, oo.

o, oo.

o, oo.

o, oo.

o, oo.

o, oo.

o, oo.

o, oo.

o, oo.

o, oo.

o, oo.

o, oo.

o, oo.

o, oo.

o, oo.

o, oo.

o, oo.

o, oo.

o, oo.

Atomic group i, j FlJ NN2,0(:N3 -1.17]NHPH,
(:OOHO,621 0H,OHO,080 Pyridine substituted at 2-position NN2.-N-(py) 0.779(:0C)1
3. -N- (p4/) 0.6970CH3, -N-
(Py) 0.5690C28P, -N-(1'V
) 0.569(i)O(:3H7,-N-(+)V
) 0.569(s)OC4Hs, -N-(py)
0.569(t)OGJ9. N-(py) 0.5
69G)10. -N-(p 0.465CONH
2, -N-(+'V) 0.78ON(CH3)2.
-N-(py) 0.356NHCH3,-N-(py
) 0.568Nococo3. -N-(py) 0
．． 5920)1. -N-(py) -0,449N
uso2pu, -N-(py) o, oo.

o, oo.

0.893 o, oo.

o, oo.

o, oo.

o, oo.

o, oo.

o, oo.

o, oo.

o, oo.

o, oo.

o, oo.

o, oo.

o, oo.

o, oo.

o, oo.

1.297 o, oo.

0.000 o, oo.

o, oo.

o, oo.

o, oo.

o, oo.

o, oo.

o, oo.

o, oo.

o, oo.

o, oo.

o, oo.

o, oo.

o, oo.

o, oo.

o, oo.

Table 4 Table 4 (Continued) Corresponding structure F i, ':', : /"r'1 Corresponding structure F i, 'j', k ρ The following blanks are as follows: 1. Distribution coefficient estimation processing method of the present invention We will explain the specific processing when implementing this using a computer.

First, a bond table is created for the target chemical substance.

A bond table is a list of how atoms are bonded by assigning a number to each atom contained in a chemical substance and expressing the types of bonds between atoms with numbers.
It is convenient and preferable to treat hydrogen atoms together with the atoms to which they are bonded, and for example, single bonds are 1, double bonds are 2, triple bonds are 3, and aromatic bonds are 4. It is expressed as

As an example, a bond table for phenol is shown in Table 5.

Table 5 Number Element bond partner number Bond type l OH21 2C137144 3CH2444 4CH3544 5CH4644 6CH5744 7CH2644 The two-dimensional or three-dimensional coordinates of each atom may also be added to the bond table, thereby displaying the bond table in a chemical structure diagram. Alternatively, it can be used for conformational analysis.

Although the chemical structure of a substance may be entered into a computer using a linear expression such as the substance name or WLN using a keyboard, it is easier for the user to assemble a chemical structure diagram using a graphics terminal. It is also convenient from At this time, a join table is created for recycle.
It is preferable that the

With this bond table, it is possible to easily recognize rings that do not contain saturated carbon, such as aromatic rings, and necessary atomic groups.

Next, based on the created bond table, for example, isolated carbon atoms are detected, fragments are cut out, and rings, atomic groups, and specific structures that do not contain saturated carbon are detected to determine the partial structure necessary for estimating the distribution coefficient. Detect.

Atomic group constants (α, ρ, σ
, F, f, δ, σ1. ES, etc.) from the separately stored atomic group constant file, the distribution coefficient is calculated according to the above equation (1). For information on how to detect partial structures and quickly extract various constants corresponding thereto from a file, please refer to Japanese Patent Application No. 62-214797 filed by the present applicant.
The methods described in this specification and Japanese Patent Application No. 62-219395 can be effectively used.

Next, output the final calculation result. Output can be performed by recording on various recording materials such as plain paper using an appropriate recording device, or by displaying (graphic display) on a color cathode ray tube or the like connected to a computer or electronic equipment.

Furthermore, the calculated value of the distribution coefficient obtained can be recorded and stored in the main memory of the computer, or via a suitable recording medium (magnetic disk, optical disk, magnetic tape, etc.), if necessary.

In addition, it is possible to compare and check with a large number of actual measured values that are stored separately, or to search for the distribution coefficient of a chemical substance using a large number of stored actual measured values and calculated values.

FIG. 1 shows a processing flowchart of an example of a computer program according to the method of the present invention. Further, FIG. 2 shows a configuration diagram of an example of a computer program according to the method of the present invention.

The method of the present invention will be further explained below using Examples.

The following examples will be described using a computer program (CP
A C: Computation or Porti
on Ca-efficients). The computer used was FAGOM N780 manufactured by Fujitsu.
The system was designed to input a chemical structural formula from a graphic terminal F9431B and output the results together with the chemical structural formula to a graphic terminal or laser printer (F6677B1). The required time for each item was approximately 40 seconds for inputting the chemical structural formula, approximately 5 seconds for calculation, and approximately 1 minute including output time.

All white below [Example 1] H When X=NO2, R=CH2Cl: logic Actual 1-E7)=0.59 MogP (estimated value) 0.13 0.355 0.345 -0
．． 03+ f CH+ 2f CH2+ f O
H+ f C0NH+ f c+0.43
0.66 -1.64 -2.71 0.0
6-0.12 -0.22 -0.26(-1
,16-2,71)-0,065-0,255-0,1
31-1,753+ ρ N02 × σP(Y)+
ρ Y × σP (NO2)−0,100,00,3
70,82 = 0.6 However, here, the π' value is the π value contribution from the substitution position to the fifth adjacent atom. That is, Table 6-2, 71-0, 12-0, 22 jljogP (actual value) IlogP (estimated p value) = -1,753. When there are multiple polar atomic groups, such as Y in this example, the α, ρ, and σ of the polar atomic group closest to the substitution position are
Let be the value of Y. In this example, the values of Y are α, ρ, and σ of the atomic group corresponding to CH20H of benzyl alcohol. That is, αY = αc u 2o ++ =
−0, O65ρ Y = ρ CH20H=
0・37δP(Y)=δP(C820H)=0
．． It becomes 0.

Table 6 shows calculation results for other derivatives along with actual measured values.

Yarn 1'1 below F3 HCl2 HBr2 (:HCIMe H2Br 1-IF2 (:82F H2I HMe2 C13 Br3 Pr e H2ON (:6H5 CH2C6H5 (:H(:N’C6H5 HEt2 Me3 1.07 1, O5 1,36 0,98 0,66 0,42 0,15 1,03 0,79 1,97 2,17 0,87 0,03 -0,22 1,63 1,52 1,47 1,71 1,33 0,84 1,08 1,31 0,99 0,75 0,34 0,17 1,14 1,03 1,80 2,02 1,19 0.08 0,05 1,57 1,57 1,72 2,11 1,41 [Example 2] Table 6 (continued) Y GO(:O3CHCl2 T CHCl2 (N OH(:1z 3-NO2CHCl2 Br CHCl2 (:I CHCl□ 1-Pr CHCl2 HCHCl.

5O2CH3C: HGIz Cr, O5CHCl2 Nl2 CHCl2 HCO- -cyclo-C3H5CHCl2 CONH2(:HCl2 NHCONH2CI((:5 uogP (actual value) 0.57 2.22 0.78 1.15 1.87 1.6 4 2.24 0 .94 0.27 2.86 -0.29 0.75 -0.55 -0.60 nogP (estimated value) 0.60 2.17 0.72 ], 04 1.95 1.71 2.34 0 .87 0.22 2.75 -0.33 0.53 0.33 -0.08 2.3, 4, 2°, 3°, 4゛For substituted compounds: sentence o
gP (actual value) = 6.93 sentence ogP (estimated value) [The above terms are the breakdown of logP additive] 0
．． 0 0.715 0.0 0.715+1
2αCIXπ'Cl -0,0670,715 [The above terms are the breakdown of Σα1 (Σπ゛'')] [The above terms are the breakdown of Σρ 1 (ΣσJ)] +fx X CI (CI) + δXXES (CI
)0.0 0.47 0.0 -0.
97+ f c+X cr I(X)+ δ c+X
E 5(X)0.0 0.12 0.0
-1.01 [The above terms are the breakdown of Fx and CI] 0.0 0.12 0.0 -1.01
Jl) [The above terms are the breakdown of FY and CI] H + 4FCI CI 0.039 = 7.18 However, here, the π' value is the π value contribution from the substitution position to the third adjacent atom. That is, = 1, 9
It was calculated as 55.

Table 7 shows calculation results for other compounds with different substitution positions and numbers of substitutions in the 0 sentence, together with actual measured values.

Replacement position 2.5 4.4' 2.6 2.4.5 2.4.6 2.3, 4.5 2.4, 5.2' 2.3, 4, 5.6 2.4, 5.2', 5' 2.3, 6.2', 3', 6° 2.4, 5.2', 4°, 5' 2.4, 6, 2°, 4°, 6° 2. 3, 4, 6, 2°, 3°.

4°, 6° 2. 3, 5, 6, 2 degrees.

3゛, 4°, 5°, 6゛ 2. 3, 4, 5, 6, 2°.

3°, 5°, 6゛ 2.3.4,5,6.2'.

3°, 4°, 5°, 6° Table 7 fl ogP (actual measurement) 4.50 4.95 5.08 5.16 5.58 4.93 5.51 5.47 5.72 5.73 6.30 5.92 (6,11) 6.63 6.72 6.34 (7,55) 6.68 7.11 8.16 8.26 JZogP (estimated value) 4.71 4.71 4. 71 5.23 5.32 5.32 5.69 5.65 6.10 6.13 6.37 6.67 6.81 7.10 6.64 7.33 7.51 7.75 8.00 [ Example 3] In the case of X = 3-N 02: ogP (actual value) = 4.57 ogP (estimated value) = 4f + 8f + f C0OH + f Nu + f
802 + F bO,130,355-0,03
-0,09-0,03-0,12[The above terms are log
Breakdown of P additive]+αcoouXπ'N+
+ +asHXπ'cooHO,01,2150,0-
0,255 +αN02x? r'mouth+αNH×π'NO2O, 131
1,2150,0-0,255 [The above terms are Σαi(Σ
Breakdown of π'))] 0.893 0.46 0.0 -0.41 0.893 0.70 -0.10
−0.15 [The above terms are a breakdown of Σρ1 (Σσ”)] H + F C00I+ 88 0.621 In this example, ρNH1ρC0OH is one of the features of the present invention ρ
, is the ρ value of NH when C0OH is present at the ortho position, and ρC0OH is the C value when NH is present at the ortho position.
This is the ρ value of 0OH.

There are two values of π NH, one corresponding to αNH and one corresponding to α802, each targeting up to the fifth neighboring atom, x = fNH + 2f; + 4fc++
fso2-f++0.09 0.13 0.3
55 -0.03 0.225+Fb -0,12 1,215 = f N, + -0,09 2fC+ 0.13 11 m 4f cs+ f C00H-f + chanting 0.355
It was calculated as -0.03 0.225+Fb O012 =1.215.

Table 8 shows the calculated values for other substitutes together with the measured values.

Less than 1・°Total 1'1 ρ is changed by incorporating the effect of proximity.

[Example 4] Table 8 When X = 4-OCH3 (7): logP (actual value) = 1.41 iogP (calculated value) uogP (actual value) 11ogP (estimated value) -CH3 -OH ■ 2.3 -di CH3 -CF3 4.88 3.49 4.36 5.12 5.25 (5,62) 4.79 3.61 4.38 5.07 5.18 0.89 -0.12 [The above The term is the breakdown of logP additive]+
α0×π N=N+αN=N Xπ100.115 −
2.337 −0.265 −0.065 [The above terms are the breakdown of Σα1 (Σπ°, )] +ρN=N Xσ−(0)
+ρ0×σ・(N=N)0.0 0.06 0.26
9 1.127 [The above terms are the breakdown of Σρ, (Σσ”)]
=1.34 In this example, fC・, fCm are both 1o of the carbon at the condensation position.
The former represents the case where the heteroatoms are not adjacent, and the latter represents the case where the heteroatoms are adjacent. In addition, a heteroatom group having an aromatic ring structure (in this example, N
=N) is calculated from the logP contribution of the heteroatom group to the logP contribution of the CH of the benzene ring to simplify the calculation.
It was calculated by subtracting the gP contribution. That is, π N-N= f N=N - f c++=
-2,3371,9820,355, again.

π o=fo-fcu3 +Fb-fu = -0
,065-0,610,89-0,120,225.

Table 9 shows the calculated values for other substitutes together with the measured values.

The following is Kinpaku Table 9 j2ogP (actual measurement value) -CI -Br -NO2 3-(1;0(:I3 3-N)I2 3-[;00C2■5,7-C2H5 1,57 1,72 0, 73 1,44 1,00 2,31 4-CI(3,6-11:1 4-CH3,7-C1 -NH2 4-Nl2,8-CH3 1,96 1,98 1,04 1,57 1ogP (Estimated value) 1.46 1.60 0.76 1.26 0.87 1.96 2.05 2.05 0.84 1.40 Below 2ζ mill [Example 5] 0.0 1.389 iogP ( Actual value) = 2.35 Sentence ogP (calculated value) +fcn+ 2fcn3+fc++ (4-1) Fb
+FcarO, 430, 890, 94-0, 12-0
, 13 [The above terms are the breakdown of logP additive]+α502N-CHNXπ C4119-is.

O,3061,97 +αc4uq-+soXπ' 5O2N-CHN-0,
073-2,71 + 0.519Xα5O2N-CHNXπ'c-0,3
060,715 + 0.519XαCIXπ'502N=CIlN-0
,067-2,71 + 0.519X a C41H9-isoXπ°C-
0,0730,715 + 0.519X αCIX π 1.4H9-is
.

-0,0671,97 [The above terms are the breakdown of Σα1 (Σπ'j)] + /) C
4H9-1soX (r p(SO2N=CNH)0.
0 1.389 [The above terms are Σρ
Breakdown of l(Σσ'')] = 2.36 As in Example 4, the π' value of the aromatic ring-configured heteroatom group (502 N=CNH in this example) is calculated as follows to simplify the calculation: It was determined by subtracting the OgP contribution of CH in the benzene ring from the JJogP contribution of the heteroatom group.

That is, calculate as follows, and for other π' values up to the 5th neighboring atom, the remaining 11 π C4119-iso = f CIH+ f
CH+ 2f CH30,4750,430,89 + 3Fb +FqBr f++0.12
-0.13 0.225= 1.97 π c+ = f c+ -f H= 0.7
It was calculated as 150.94 0.225. In addition, C1 and C4Hq-rso, Cl
Since 5OyN=CNH are atomic groups substituted on separate adjacent rings, as described in (F), the term representing the interaction was multiplied by 0.519.

Table 1O shows the calculated values for other substituents together with the measured values.

■ -CH3 3-CH.

-CH3 -CH3 -C13 -CH3 3-(:O3 3-(:O3 3-(:O3 -CH3 1: Margin -NHAc -C113 3-(:O3 -C13 -CH3 -CH3 -C2H5 -CH3 10th Table λogP (actual value) It ogP (estimated value) ■ -CH3 5-Br-7-CI -Br 5-1-7-CI 5-NO2-7-CI -10H3 -Br -CI -CI -7-502NH2 -CI 6-NO2-7-(:1 -0CH3 6- to F3 6-NH2-7-CI -CJ5 -C2H5 7-B' 0.16 (0,18) 0.29 0.52 1.65 0 .72 1.81 0.85 0.74 1.37 1.21 0.03 0.17 0.34 0.42 1.65 0.79 1.82 0.90 0.80 1.35 1.14 0.14 0.53 1.42 0.56 1.59 0.63 1.25 0.81 1.37 0.30 1.24 0.50 1.54 0.73 1.17 0.80 1. 35 Table 10 (continued) Table 10 (continued) II ogP (actual value) 1 ogP (estimated value) Standing ogP (actual value) fl ogP (estimated value) 3-(:)13 -CH3 -CH3 3-( :H3 3-C:H37 -CH3 -F -CI -NO2 SO3M (CH3) 2 -CF3 0.62 1.20 (1,08) 1.61 0.65 0.32 1.51 0.55 1. 14 1.55 0.69 0.43 1.54 3-(:(CH3)+ -CF3 3-iso-Pr 3-CHC:12 3-sec-Bu 7-502NH2 7-(:1 -C1 -C1 -C1 -C1 1,40 2,40 1,65 2,00 1,81 2,47 1,46 2,31 2,04 2,01 1,80 2,36 +-y+ 3-CH2(:l 6 -C1 3-C12001136-C1 3-benzyl H 1,68 1,12 1,89 1,55 1,01 1,89 -y1 -3-y+ norbornen H 2-yl 3-phenyl H 2,06 1,78 1,78 1,78 3-Et-5(1:H26-C1 3-propenyl 6-C1 3-neopentyl 7-C1 3-(:2H56-(:1 3-C2H57-CI 3-C3H76-11:1 2.26 1.96 2.71 1.61 1.62 2.04 2.19 2.12 2.63 1.83 1.83 2.18 Table 10 (continued) [Example 6] j2ogP (actual measurement Value) 3-Bu 7-CI 3-C:J++ 6-Gl- 7-5O2NH2 H6-CH3- 7-5021u12 H6-Br- 7-502NH2 H6-F- 7-502NH2 H6-C1 3-cycloPr 6-CI H6-C1 7-502NH2 H6-N02 7-5O2NHz 7-5O2NH2 H7-(:1 3-cycloBu 7-CI H7-5O2NH2 2,52 2,08 -0,32 0,00 0,29 1,02(1 ,01) 1.98 -0.27 0.08 1.05 1.22 (1,11) 2.24 0.62 Il, ogP (estimated value) 2.54 1.78 0.12 0.03 - 0.29 1.02 1.77 -0.02 0.08 1.05 1.02 2.15 -0.58 In the case of X = NH2: Also, ogP (actual value) = -1,09 1agP (calculated value) [The above terms are the breakdown of logP addiLive] +
αNH2Xπ'NllN=CH+αN118−CHXπ
″NH2-0,075-0,8250,0-1,225
[The above terms are the breakdown of ΣαI (Σπ'j)] +ρ NHN
=CHX σHatake(11)12)+ρ Nl2 X
σ (NHN = CH) 0.78 -0.1
4 0.67 0.134 [The above terms are Σρl(
Σσ)) = -0, 94 Similarly to Example 4.5, the π° value of the aromatic ring-configured heteroatom group (NHN=CH in this example) is as follows to simplify the calculation: It was determined by subtracting the logP contribution of CH in the benzene ring from the lagP contribution of the heteroatom group. In other words, it was calculated as follows.

Table 11 shows the calculated values for other substitutes together with the actually measured values [Example 7].

11th table ogP (actual value) lagP (calculated value) 5H11 ■ O2 0M NHGOC:H:+ NHSO2-φ CH3 2,96 1,70 0,59 0,24 -0,38 0,87 0,37 2 ,89 1,75 0,63 0,25 -0,38 0,92 0,35 Below margin Nl2 fLogP (actual value) = -1,09 fLogP (calculated value) +2f Nl2 1.00 [The above terms are l Breakdown of ogP additive] +
2XαN Xπ's+ 4X(XN Xπ'Hn2−
0.178 -1.475 0.178 -1.225
+ + + 4×αN112Xπ'%+ 2X αN82 X
π'NH2-0.075-1.475-0
．． 075 -1.225 [The following terms are Σα1(Σπ
°〕）〕2×ρ=H arrow σ−(N)+4X
ρ = H2 × σ9 (N) 0.0 0.596
0.0 0.7254Xρ = Takashi σp(882)+
2X ρ; H2X σ-(882)0.0-
0.38 0.0 -0.14 [The above terms are the breakdown of Σρl (ΣσJ)] + 3F N +u2
+ F N, Nl2. No. 779-0.
822 [The above terms are the breakdown of ΣF 1 = 0.91 In this example, Σαl (Σπ“”) + Σρ, (Σσ, ) +
Since the total correction of ΣF is 3.538, which exceeds the correction upper limit of 2.8, this part was calculated as 2.8.

Table 12 shows the calculated values for other derivatives together with the measured values.

Margin below Table 12 Substituent X ILogP (actual value) -Or -NH2 4-C:O3 -0,40 0,72 -0,22 0,16 4,6-diCH3 4-Nl2,6-CH3 2,6-dicH3,4-Nl2 O,62 0,19 0,39 1ogP (estimated value) -0,30 0,71 -0,13 0,23 0,66 0,34 0,77 4-φ 2-Nl2,4-φ 2-NHCH3,4-φ 2-011:O3,4-φ 2-(:O3,4-φ 2.71 2.95 3.58 3.50 3.10 -N -N 2.61 2.92 -Nl2 -0CH3 -C1 1,28 1,97 1,74 2,52 2,58 3,43 3,50 2,88 2,77 2,50 1,29 1,95 1,77 The * mark applies the correction upper limit value of 2.80.

[Example 8] [The above terms are the breakdown of ΣF] Table 13 shows the actual measurement of calculation i1 for other derivatives
2-C sentence, 4.6-diN HC2Hs: Show with values.

Sentence ogP (actual value) = 2.18 1ogPc calculated value) Below 0.89 -0.12 [The above terms are the breakdown of logP additive] 6Xα) I X
π'11+ 3X(XN Xπ'C-0,178-1,
475-0,1780,7150,0-1,475 [The above terms are Σα, breakdown of (Σπ'J)] 0.0 0.27 + 6X ρ H 0,0 X σp(NHC2)15) -0 , 41 [The above terms are the breakdown of Σρ l (Σσ'')] [Example 9] Table 13 Substituent X IlogP (actual value) uogP (estimated value) OH -0,73 0,72 2-CI.

4.6-diN(CH3)2 2.312.4.6-t
riN(CH3) 22.732.31 2.65* Bun ogP (actual value) = 1.05 Bun ogP (calculated value) 4fμ + 2fS+f C0OH+ 3f onO
, 130, 355-0, 03-0, 44 [The above terms are l
Breakdown of ogP additive] + 3× α
C00HX π 1OH0,0-0,665 [The above terms are the breakdown of Σα1 (Σπ'j)] 0.04
−0.13 [The above terms are the breakdown of Σρl (Σσ”)] + F OIl
, C0OH+ 2F o++ OH1,0070,
08 [The above terms are the breakdown of ΣF] = 1.05 Table 14 shows calculated values for other benzene derivatives along with actually measured values.

As is clear from the examples below, the distribution coefficients calculated using the estimation method of the present invention reproduce well the actual measurements for all compounds. It can be seen that this is an estimation processing method for making good predictions.

[Brief explanation of the drawing]

FIG. 1 is a processing flow diagram of an example computer program according to the method of the present invention. FIG. 2 is a block diagram of an example of a computer program according to the method of the present invention. Patent applicant Fuji Photo Film Co., Ltd. Representative
Patent Attorney Yasushi Yanagawa Male 11th Me1

Claims (1)

[Claims] 1. A method for estimating the distribution coefficient of a chemical substance having a ring that does not contain saturated carbon; Five atomic group constants; of the amount of change π in the logarithm of the distribution coefficient when one atomic group is substituted in an aromatic ring, the distance from the substitution position is the number of bonds (n is the number of bonds (however, n is 1≦ n≦
contribution by the substructure consisting of atoms in the range of 10 (an integer) π', the measure α by which one atomic group changes the π' value of another atomic group
, the electronic substituent constant σ of one atomic group, the measure ρ by which the π value of one atomic group changes depending on the σ value of the other atomic group, and the substitution of two or more atomic groups in close proximity to each other. Given the amount of change F in the logarithm of the partition coefficient when It is the distribution coefficient of a substance, and logP_a_d_d_i_t_i_v_e is based on the sum of the contributions of the logarithm of the distribution coefficient assigned to each atomic group constituting the substance and the proximity of bonds, branches, and polar groups in the structure of each atomic group. is the total sum of the correction factors,
(i and j each represent an atomic group contained in the substance) and calculating a distribution coefficient of a chemical substance using the formula (1). 2. Distribution of chemical substances according to claim 1, characterized in that the atomic group constant ρ_i for the atomic group i can be changed depending on other atomic groups j in the vicinity of the atomic group i. Coefficient estimation processing method.