J. Chil. Chem. Soc., 55, Nº 3 (2010)
Physicochemical significances of the proposed models
CONCLUSIONS
It will be interesting to discuss the physicochemical significances of the
proposed models. We observed that invariably all the models contain C2
(∑NMR) as the correlating parameter. It means, therefore, that ∑NMR play
a dominating role in modeling antibacterial activity against all the bacteria.
That is electronic effect is the main parameter for the exhibition of the activity.
In case of models -3 and model-5, in addition to ∑NMR, physicochemical
parameters along with topological indices are involved in modeling the
antibacterial activities. The physicochemical parameters involved indicate
that the size and shape are responsible for the exhibition o the antibacterial
activities against these bacteria. The topological indices involved in these
and other models are mainly connectivity indices. This means that the extent
of connectivity is the responsible for the exhibition of anti bacterial activity
against all the bacteria used.
The newly synthesized Mannich bases appeared to be very potent and
outstanding antibacterial agents with promising activity and found safer.
These novel Mannich bases could be used as useful drug. Our findings will
prove helpful to those who are engrossed in the synthesis of potential Mannich
bases as drugs with minimum side effects and also having comparatively
low cost. Thus, the result presented in this paper is valuable in constructing
pharmacologically imperative heterocyclic as new exotic drugs. Efforts are
continuing to synthesize new amino-methyl derivatives of various active
hydrogen compounds, that the derived compounds may have enhanced
pharmacological activity. Our results also show that the antibacterial activity
of the Mannich bases could be modeled using sum of the NMR chemical shifts,
physicochemical parameters, and topological indices. The combination of these
parameters gives statistically significant models for modeling antibacterial
activity against S.typhi, E.coli, S.aureus, and K.pneumoniae.
It is also worth mentioning that all the proposed models, except model-5
(which has 7 correlating parameters), all other models contain 8 correlating
parameters. Looking to the size of the sample i.e. the number of compounds
used this is accordance with the rule of thumb. Furthermore, the plot of number
of variables used against R2
ACKNOwLEDGEMENTS
( Fig.1) also favors the use of 7 to 8 correlating parameters.
The aforementioned models were further examined employing Ridge
statistics.
The authors wish to express gratitude to the Director, CDRI, Lucknow,
India for recording elemental analysis.
REFERENCES
Ridge statistics.
Application of Ridge statistics provides important statistical parameters
namely variance inflation factors (VIFs) for each of the parameters involved in
the model. The VIF is defined for each variable in the equation, and not for the
equation as a whole, so there should be as many VIFs, as there are correlating
parameters. The VIF is defined as:
1. C. Hansch, A. Leo, In Exploring QSAR: Fundamental and application
in chemistry and biology ,S.R. Heller, Ed. American Chemical Society:
Washington, DC. (1995)
2. M. Karelson, Molecular descriptors in QSAR/QSPR, Wiely-Interscience
(2000)
3. M.V. Diudea, M.S. Florescn, P.V. Khadikar, Molecular topology and its
applications, EFICON Press: Bucarest. (2006)
VIF = 1(1-R2i)
(19)
4. C. Hansch, B.R. Telzer, L. Zhang, Crit. Rev. Toxicol. 25,67- 89. (1995)
5. H. Branes, Mamalian toxicology-property prediction and QSAR
technique,Peter Fisk Associates Whitestable: Kent. (2004)
6. S. Bradbury, Taxicol. Lett. 79,229-237, (1995)
7. P.R. Duchowicz , E.A. Castro, A.A. Toropov, E. Benfenati, Top
Heterocycl. Chem 3, 1-18 , (2006)
8. M.V. Diudea, (Ed) QSPR/QSAR Studies by Molecular Descriptors, Nova
Science, (2000)
9. J. Devillers, A.T. Balaban, (Eds) Topological Indices and Related
Descriptors in QSAR and QSPR, Gordon & Breach Science Pub.,
Amsterdam (The Netherlands), (1999)
Where Ri is the multiple correlation coefficient of the ith independent
variable on all of the other independent variables. In the proposed models, all
these VIFs should be less than 10 indicating that no co- linearity problem exists
in the model.
The VIFs values for the parameters involved in the aforementioned models
are given supplementary material; which shows that each of the models contain
one or more descriptors having VIFs values larger than 10. Therefore, based
on VIFs values a major problem of co- linearity exists for them. However, the
magnitude of condition number indicates the existing of mild co- linearity. In
such cases Randic recommendations 59-60 are used to arrive at the final decision.
10. M.V. Diudea, (Ed) QSPR/QSAR Studies by Molecular Descriptors, Nova
Science( 2000 )
Randic recommendations
59-60
Randic
stated that if a descriptor strongly correlates with another
11. D. Janezic, Nikolic, T.N.Sonja, Graph theoretical matrices in chemistry,
University of Kragujevac, Faculty of Science, Kragujevac,205 ( 2007)
12. Li, Xueliang , Tianjin, I.Gutman, Mathematical Aspects of Randic-
Type Molecular Structure Descriptors Nankai University, University
of Kragujevac, Kragujevac, Serbia University of Kragujevac,
Faculty of Science, Kragujevac ISBN: 86-81829-67-X, 330 ,(2006)
Rouvray , H. Dennis (Editor), Complexity: Introduction and Fundamentals
University of Georgia, Athens, Georgia, USADanail Bonchev (Editor),
Virginia Commonwealth Univ., ichmond, USACRC PressISBN: 0-415-
28791-X, 208, (2003)
descriptor already used in a regression, such a descriptor in most studies
should be discarded. For example χ and χ , 1χ often strongly correlate and
1
2
2
in many structure-property-activity studies χ has been discarded. This is not
theoretically justified and despite the widespread practice should be stopped.
Although two highly correlated descriptors overall depict the same features of
molecular structure, it is important to recognize that even highly interrelated
descriptors differ in some other structural traits. The difference between them
may be relatively small but nevertheless very important for structure-property
regression.
The criteria for inclusion or exclusion of descriptors should not be based
on parallelism between descriptors even if overwhelming, but should be based
on whether the part in which two descriptors disagree is or is not relevant for
the characterization of the property considered .If the part in which the second
descriptor differ from the first, regardless of how small it is, is relevant for the
property under consideration, then the descriptor should be included. Randic 59,
60 further stated that the selection of descriptors to be used in structure-property-
activity studies should not be delegated solely to computers, although statistical
criteria will continue to be useful for preliminary screening of descriptors taken
from a large pool. Often in an automated selection of descriptors, a descriptor
will be discarded because it is highly correlated with another descriptor already
selected. But what is important is not whether two descriptors parallel one
another; i. e. duplicates much of the same structural information, but whether
they are complementary in those parts that are important for structure-property-
activity correlations. Hence, the residual of the correlation between two
descriptors should be examined and kept or discarded depending on how well
it can improve the correlation based on already selected descriptors.
13. R. King, Bruce (Editor), Topology in Chemistry: Discrete Mathematics
of Molecules University of Georgia, Athens, Georgia, USADennis H.
Rouvray (Editor), University of Georgia, Athens, Georgia, USAALBION/
HORWOOD Pub.ISBN: 1898563764, 400 ,( 2002)
14. H.Diudea,
V
Mircea., I.Gutman, Molecular Topology NOVA
SCIENCEISBN, 332 , (2001)
15. R. Todeschini, Handbook of Molecular Descriptors WILEY-VCHISBN,
668, (2000)
16. L.Kier, R.Hall, Molecular Structure description ACADEMIC PRESS
ISBN,286, (1999)
17. A.T.Balaban , (Editor), From Chemical Topology to Three-Dimensional
Geometry PLENUM PRESSISBN, 420 ,( 1997)
18. G.P.Mezey, Shape in Chemistry: An Introduction to Molecular Shape and
Topology JOHN WILEY & SONS, 224 ,(1993)
19. N.Trinajstic, Chemical Graph Theory CRC Press, 322 ,( 1992)
20. H.D.Rouvray , (Editor), Chemical Graph Theory: Introduction and
Fundamentals USATAYLOR & FRANCIS, 300 ,( 1991)
21. R.B.King , (Editor), Graph Theory and Topology in Chemistry, USA
ELSEVIER, 575 ,( 1987)
287