Martins J. P. A., Barbosa E. G., Pasqualoto K. F. M., Ferreira M. M. C., "LQTA-QSAR: a new 4D-QSAR methodology". Istanbul, Turkey, 04-08/09/2009: Fifth International Symposium on Computational Methods in Toxicology and Pharmacology Integrating Internet Resources (CMTPI 2009), Abstract Book (2009) 56. Oral OC-12.
LQTA-QSAR: A NEW 4D-QSAR METHODOLOGY
João Paulo A. Martins, Euzébio G. Barbosa, Kerly F. M. Pasqualoto, Márcia M. C. Ferreira*
Laboratory for Theoretical
and Applied Chemometrics, Institute of Chemistry, University of Campinas,
Campinas, SP 13084-971, Brazil, *e-mail: marcia@iqm.unicamp.br
The new 4D-QSAR approach
presented and named LQTA-QSAR1 (Laboratório
de Quimiometria Teórica e Aplicada), is based on the generation
of a conformational ensemble profile, CEP, for each compound, followed
by calculation of 3D descriptors. This new methodology explores jointly
the main features of CoMFA and 4D-QSAR paradigms. GROMACS free package
is used for molecular dynamics, MD, simulations and generating CEP. The
module LQTAgrid calculated intermolecular interaction energies at each
grid point considering different probes and all aligned conformations from
MD simulations. These interaction energies are the descriptors employed
in the QSAR analysis. The ordered predictor selection, OPS, algorithm2
recently developed in our laboratory, is applied as the variable slection
method in the construction of the PLS models. OPS method has been proved
to be fast and capable of providing suitalble variables fro the QSAR analysis.
LQTA-QSAR models are thoroughly validated applying the leave-N-out
cross-validation and y-randomization methods. The comparison of the proposed
methodology to other 4D-QSAR and CoMFA formalisms was performed using a
set of forty-seven glycone phosporylase b inhibitors (data set 1) and a
set of forty-four MAP p38 kinase inhibitors (data set 2). The QSAR models
were built using the OPS algorithm for variable selection. Model validation
was carried out applying y-randomization and leave-N-out cross-validation
in addition to the external validation. PLS models for data sets 1 and
2 provided the following statistics: q2
= 0.72, r2 = 0.81 for 12 variables
selected and 2 latent variables; and, q2
= 0.82, r2 = 0.90 for 10 variables selected
and 4 latent variables, respectively. Visualization of the descriptors
in the 3D space was successfully interpreted from the chemical point of
view, supporting the applicability of this new approach in rational drug
design. LQTA-QSAR is availabel at http://lqta.iqm.unicamp.br
References
1Martins JP;
Barbosa E; Pasqualoto KF; Ferreira MMC, LQTA-QSAR: a new 4D-QSAR methodology.
J. Chem. Inf. Mod. 2009, in press.
2Teófilo
RF; Martins JP; Ferreira MMC, Sorting variables by using informative vectors
as a strategy for feature selection in multivariate regression. J. Chemometr.,
2009, 23, 32.
Acknowledgements:
FAPES, CNPq, CAPES