Você está na página 1de 11


In present study, methodology is based on aspect of Quantitative Structure Activity Relationship i.e, to develop mathematical model based on relation: = f(C) where, = Biological property C = Structural descriptor/ physicochemical properties C used in present work is topological indices and physicochemical properties. The following topological indices are tested in present study: a. Wiener Index (W)

Randic` Connectivity Index ( )

c. Balaban . J. Index (J) d. Szeged Index (Sz) e. Molecular Topological Index (MTI) f. Electrotopological Index (S) a: The Wiener Topological index: Wiener index 21 is defined as the sum of the chemical bonds existing between all pairs of carbon atoms in the molecule under consideration. The index was later shown by Hosoyo to be equivalent to one half the sum of the entries in the distance matrix of the graph of the molecule. In symbol, the Wiener index can thus be defined as follow: W(G) = 1/2 dij i j Where dij represents the ijth entry in the distance matrix of the graph G.

Example: 2,2,3trimethyl butane 6 1 2 5


3 7

01 10 21 32 21 21 32

2 1 0 1 2 2 1

3 2 1 0 3 3 2

2 1 2 3 0 2 3

2 1 2 3 2 0 3

3 2 1 2 3 3 0

WIENER INDEX W(G) = 42 b: Topological index of Randi'c or Connectivity index. The next major topological index was that of Randi'c28. This index, which is nowadays widely referred as the molecular connectivity index. The index has the symbol and is as widely known in the biological sciences and in the physical sciences because of its widespread application. In its original form index was defined in the following way: (G) = (Pi Pj)-1/2 edges where Pi and Pj represents the degree of the adjacent pair of the vertices i and j in G for graph(G) of 2,2,3trimethyl butane is calculated as follow. (G) = 1/1x4 + 1/4x3 + 1/3x1 + 1/1x4 + 1/4x1 + 1/3x1 = 2.94338

c: The Topological index of Balaban The topological index of Balaban17 is based on the distance matrix of the graph G and is known as the average distance sum connectivity index. For tree graph The index may be written as J(G) = ne x (Si Sj)-1/2 Where ne is the number of edges in the tree and Si and Sj represents, respectively, the sums of the ith and jth row of the distance matrix of the tree. Example 2,2,3trimethyl butane Distance matrix of graph (G) 2,2,3trimethyl butane 0 1 2 3 2 2 3 1 0 1 2 1 1 2 2 1 0 1 2 2 1 3 2 1 0 3 3 2 2 1 2 3 0 2 3 2 1 2 3 2 0 3 3 2 1 2 3 3 0 = 12 =8 =9 = 14 = 13 = 13 = 14

J(G) = 6x (1/12x8 + 1/8x9 + 1/9x14 + 1/13x8 + 1/13x8 + 1/14x9) = 3.5412 The index attempt to reflect both molecular size and the extent of branching present and increases with both of these parameters. d: The Szeged index Another molecular-graph-distance-based topological index namely Szeged index has been used in the present investigation. This index was conceived in Szeged Hungary by Gutman et al153. The Szeged index, Sz = Sz(G) of G is defined as

Sz = Sz(G) = nunv

where summation goes over all edges (u,v) in a cyclic graph G, Where, nu and nv are the number of vertices on either side of the vertex under consideration. Example for a graph (G) of 2,2,3trimethyl butane Sz(G) = [(6x1) + (4x3) + (6x1) + (6x1) + (6x1) + (6x1)] Sz(G) = 42 e: Molecular Topological Index: Molecular topological index was developed by Harry. P.Schultz 154 by the use of the Adjacency, degree and distance matrices that describe the structure of compounds.The MTI of graph(G) can be calculated as MTI(G) = v(D + A) Where, v = valency degree matrix D = Distance matrix A = Adjacency matrix Example MTI for Graph of 2,2,3trimethyl butane is calculated as follows. v(D+A) 0 2 2 3 2 2 3 2 0 2 2 2 2 2 2 2 0 2 2 2 2 [ 1 4 3 1 1 1 1] 3 2 2 0 3 3 2 2 2 2 3 0 2 3 2 2 2 3 2 0 3 3 2 2 2 3 3 0 v(D+A) = [24 + 16 + 18 + 24 + 24 + 25] = 127 MTI(G) = 127

f: Elecrotopological Index Electrotopological index was developed by L.B Kier and L.H Hall13. It can be calculated by the use of Intrinsic value of each group. The Intrinsic state of an atom in a chemical graph reflects its electronic and topological attributes in the absence of interaction with the rest of the molecule. The state of each atom in a chemical graph due to the intrinsic state of that atom and the molecular field may be called the electrotopological state, S. It is caculated to be the sum of these two terms S = Ii + Ii Calculation of Electrotopological state for a graph(G) of 2,2,3trimethyl butane Intrinsic values I1 = 2 I2 = 1.25 I3 = 1.33 I4 = 2 I5 = 2 I6 = 2 I7 = 2 (Ii-Ij/r2) marix j i 1 0.0000 2 -0.1875 3 -0.0744 4 0.0000 5 0.0000 6 0.0000 0.1875 0.0200 0.0833 0.1875 0.1875 0.0744 0.0000 0.0000 0.0000 0.0000 0.0000 -0.0200 -0.0833 -0.1875 -0.1875 -0.0833 0.0000 -0.1675 -0.0744 -0.0744 -0.1675 0.1675 0.0000 0.0000 0.0744 0.0000 0.0000 0.0744 0.0000 0.0000 0.0000 0.0000 0.0000 1 2 3 4 5 6 7 I= row sum 0.2619 -0.7491 -0.5382 0.2508 0.2619 0.2619

0.0000 0.0000 0.0000

7 0.0000


0.1675 0.0000

0.0000 0.0000 0.0000

0.2508 0.0000

S i = I i + Ii S1 = 2.2619 S2 = 0.5009 S3 = 0.7918 S4 = 2.2508 S5 = 2.2619 S6 = 2.2619 S7 = 2.2508 S(G) = Si S(G) = 12.58 All the above indices are calculated using computer program MOLTOP155 g: Physicochemical properties used in present investigation are The physicochemical properties describe various structural, physical & chemical assets of the compounds viz., size, polarizability, membrane transportation, inter and intra molecular forces vander Waals volume, weight etc. these physicochemical features plays the dominating role in deciding the biological activity or function of any molecule or chemical systems. The dependence for CNS depressant activity of the molecules derived from Benzodiazepines over the physicochemical properties are proposed to study in the present work is as follows


Molar Refractivity (MR)

The molar refractivity is the molar volume corrected by the refractive index. It represents size and polarizability of a fragment or molecule. MR = n2-1 / n2 + 2 * M.Wt/D where,

n = index of refraction

2. Molar Volume (MV) Molar volume (MV) is a polarizability parameter represented by MV = M.Wt/D Where, M.Wt = Molecular weight D = Density 3. Parachor (Pc) Parachor (Pc) is an additive constitutive property of a molecule and is related to the molar volume and the surface tension. The parachor of benzodiazepines can be calculated from its constituent atoms and bonds. The parachor of a biologically active molecule is related to the ability of that molecule to permeate hydrophobic regions of cells, especially cellular membranes. An examination of the parachor values of a large number of compounds shows that these values are correlated with a number of different biological activities, PA =

* M.Wt/D

= Surface Tension

Index of Refraction ()

The refractive index, n, of a chemical system is defined as the ratio of the velocity, c, of a wave phenomenon such as light or sound in a reference Chemical system (medium) to the phase velocity, vp in the medium itself:

It is most commonly used in the context of light with vacuum as a reference medium, although historically other reference media (e.g. air at a standardized pressure and temperature) have been common. It is usually given the symbol n.

5. Surface Tension (ST) The parameters accounted for the intra and intermolecular forces and steric effect definitely plays a significant role in the modeling of biological activity of the compounds Surface tension (ST) or Inter facial tension is the cumulative effect of the different intra and intermolecular forces of two different surfaces. ST = (Pc/MV)4 Since ST is inversely proportional to MV, therefore, this is inverse steric effect. 6. Density (D) Density (D) is the microscopic density of the molecule (molecular weight / Van der Waal's volume). This may exhibits the significance of vander Waals volume in respect to weight for bio-toxic activity of the chemical systems. 7. Polarizability (Pol) Atomic polarizability is a sum over all atoms in the molecule and describing the molecule's ability to polarize in a magnetic field. Generally it is found that the more polarizable molecules are more toxic and by the use of this property we will try to specify the role of polarizability in terms of benzodiazepines. 8. logarithm of Octanol/Water Partition Coefficient (logP) The Octanol Water partition coefficient is the ratio of concentration of specific compound in an Octanol/Water mixture. log P = Concentration of a compound in Octanol Concentration of a compound in water It shows hydrophobic interaction between a drug and a binding site at a receptor. It is used as a predictor of solute-membrane partitioning.

All the above physicochemical properties are calculated using Chemsketch5.0 (www.acdlabs.com) while logP is calculated using available computer program. Indicator Parameters: Indicator parameters are also used along with indices and physicochemical properties to show the significance of substituents on various positions. Indicator parameters are descriptor having value 1 or 0 to indicate presence or absence of substituents at specific position. Regression Analysis: Regression analysis is one of the most commonly used analytical methods in chemistry and drug design. In its simpler form, regression analysis involves fitting a straight line through a set of data points represented by just two variables, calculating an equation foe the fitted line, and providing estimates of how well the points fit the line. The equation for simple linear equation is given by: y = mx + c Eq (6) the value of c, the intercept of line, can be read from the graph where x = 0 (y = m0+c) and the value of m, the slope of the line, by taking the ratio of the differences in the y and x values at two points on the line (y 2 y1)/(x2 x1). The best known regression statistics is the correlation coefficient (r2) is given by division of the explained sum of squares of the total sum of squares r2 = ESS/TSS Eq (7) This can take a value of 0, where the regression is explaining none of the variance in the data, up to a value of 1 where the regression explains all of the variance in the set, r2 multiplied by 100 gives the percentage of variance in the data set explained by the regression equation. The correlation coefficient can take values between 1, a perfect negative correlation (y decreases as x increases), and +1, a perfect positive correlation. Correlation coefficients, both simple and multiple (where several variables are involve), can be very misleading. One more statistics term standard errors gives information about goodness of fit, large standard error suggest misfit where as smaller standard error suggests good fit among the variables.


Multiple Linear Regression (MLR): MLR is an extension of simple linear regression by the inclusion of the extra independent variables Y= ax1 + bx2 + .+ constant. Eq (8) Least squares can be used to estimate the regression coefficients (a, b, c, and so on ) for the independent variables (x1, x2, x3, and so on), and the value of the constant term. Goodness of fit of the equation to the data can be obtained by calculation of a multiple correlation coefficient (r2) just as for simple linear regression. In simple linear regression it is easy to see what the fitting procedure is doing, but in case of multiple regression it is difficult to check a multiple regression fit by by plotting a data points with respect to all of the parameter in the equation, but one way that even the most complicated regression model can be evaluated is by plotting predicted y values against the observed values. A guide to over all significance of a regression model can be obtained by calculation of a quantity called F statistics. This is simply the ratio of the explained mean square (MSE) to the residual mean square(MSR) F = MSE/MSR Eq (9)

As might be expected, the F values are greater for higher level of significance The most familiar standard approaches to QSAR139 are based on multiple linear regression (MLR) and partial least squares (PLS) regression140. However, these approaches can capture only linear relationships between molecular characteristics and structural or functional features to be predicted. The major models for structure-propertyactivity studies may be grouped into the four types: empirical model; mathematical (calculus) model; mathematical (discrete) model; and graphical model (computer graphics). All these models have a common goal, which is the representation and characterization of chemical structures. They differ in their input and their output information, that is, the way in which they represent chemical structures and what quantitative descriptors they extract from their model. The ultimate criterion for validation of any of the models is the predictability, that is the potential to characterize structures and properties that were not used in modeling.


In present study linear mathematical models are developed to study Quantitative structure/Property Activity Relationship. Multiple linear regressions is used to developed these models. Topological indices and physicochemical properties are used as independent variables to predict biological properties (dependent variable). Univariate, bivariate to multivariate regression has been performed for finding out the best correlation models. All those correlation having value of R below 0.50 are considered to be statistically insignificant.