Self-organizing Neural Networks for Modeling Robust 3D and 4D QSAR: Application to Dihydrofolate Reductase Inhibitors
Abstract
:Introduction
Results and Discussion
No. | R1 | R2 | R3 | R4 | R5 | R6 | R7 | log(1/I50) |
---|---|---|---|---|---|---|---|---|
1 | OCH3 | OCH3 | OCH3 | H | H | H | H | 8.23 |
2 | OCH3 | OCH3 | OCH3 | CH3 | H | H | H | 5.85 |
3R | OCH3 | OCH3 | OCH3 | H | OH | CH3 | H | 4.00 |
4S | OCH3 | OCH3 | OCH3 | H | OH | CH3 | H | 4.00 |
5 | OCH3 | OCH3 | OCH3 | H | =CH2 | H | 5.60 | |
6R | OCH3 | OCH3 | OCH3 | H | H | CH3 | H | 5.35 |
7S | OCH3 | OCH3 | OCH3 | H | H | CH3 | H | 5.35 |
8 | OCH3 | Br | OCH3 | H | H | H | H | 8.53 |
9 | OCH3 | OH | OCH3 | H | H | H | H | 7.96 |
10 | OCH3 | OH | OCH3 | H | H | H | CH3 | 6.52 |
11 | OCH3 | OCH3 | OCH3 | H | H | H | CH3 | 7.00 |
12 | OH | H | OH | H | H | H | H | 2.78 |
13 | H | H | H | H | H | H | H | 5.71 |
14 | CH2OH | H | CH3OH | H | H | H | H | 5.83 |
15 | H | H | Cl | H | H | H | H | 6.14 |
16 | H | Br | H | H | H | H | H | 6.30 |
17 | OCH3 | H | H | H | H | H | H | 6.40 |
18 | OCH3 | H | OCH3 | H | H | H | H | 7.75 |
19 | CH3 | H | CH3 | H | H | H | H | 7.45 |
20 | H | C6H5 | H | H | H | H | H | 6.40 |
Conclusions
b | c | |||
CoMSA | ||||
all b) | MDa) | 0.5 | 0.5 | 0.5 |
q2 | 0.62 | 0.72 | 0.64 | |
S | 1.17 | 1.01 | 1.08 | |
Training/test set | MD | 0.5 | 0.5 | 0.5 |
q2 | 0.59 | 0.64 | 0.71 | |
S | 1.02 | 0.91 | 0.90 | |
SDEP | 1.25 | 0.96 | 1.20 | |
IVE | max Ac) | 6 | 5 | 6 |
MD | 0.5 | 0.5 | 0.5 | |
q2 | 0.62 | 0.87 | 0.71 | |
S | 0.89 | 0.58 | 0.79 | |
SDEP | 0.81 | 0.72 | 0.78 | |
s-CoMSA | ||||
all | sector size | 1 | 1 | 3 |
q2 | 0.38 | 0.56 | 0.47 | |
s | 1.50 | 0.90 | 1.01 | |
Training/test set | sector size | 1 | 1 | 1 |
q2 | 0.54 | 0.70 | 0.69 | |
s | 1.96 | 1.59 | 1.26 | |
SDEP | 1.42 | 1.32 | 1.42 | |
IVE | sector size | 1 | 1 | 4 |
max A | 4 | 1 | 3 | |
q2 | 0.73 | 0.59 | 0.68 | |
s | 0.83 | 0.92 | 0.85 | |
SDEP | 0.93 | 1.15 | 1.10 |
Acknowledgments
Experimental
Model builders
Kohonen mapping
Comparative Kohonen mapping
4D QSAR calculation
Calculation of the molecular surface (s-COMSA) descriptors based on virtual cubic grid
PLS analysis
References and Notes
- Kolb, H.C.; Finn, M.G.; Sharpless, K.B. Click Chemistry: Diverse Chemical Function from a Few Good Reactions. Angew. Chem., Int. Ed. Engl. 2001, 40, 2004–2021. [Google Scholar]
- Buden, F.R.; Winkler, D.A. Robust QSAR models using Bayesian regularized neural networks. J. Med. Chem. 1999, 42, 3183–3187. [Google Scholar] [PubMed]
- Schneider, G.; Wrede, P. Artificial neural networks for computer-based molecular design. Prog. Biophys. Mol. Biol. 1998, 70, 175–222. [Google Scholar] [PubMed]
- Polanski, J. Molecular shape analysis. In Handbook of Chemoinformatics; Gasteiger, J., Ed.; Wiley-VCH Verlag: Weinheim, 2003; pp. 302–319. [Google Scholar]
- Anzali, S.; Gasteiger, J.; Holzgrabe, U.; Polanski, J.; Teckentrup, A.; Wagener, M. The use of self-organizing neural networks in drug design. Perspect. Drug Discov. Design 1998, 9/10/11, 273–299. [Google Scholar]
- Lucic, B.; Trinajstic, N. Multivariate Regression Outperforms Several Robust Architectures of Neural Networks in QSAR Modeling. J. Chem. Inf. Comput. Sci. 1999, 39, 121–132. [Google Scholar]
- Lucic, B.; Trinajstic, N.; Sild, S.; Karelson, M.; Katritzky, A. R. A New Efficient Approach for Variable Selection Based on Multiregression: Prediction of Gas Chromatographic Retention Times and Response Factors. J. Chem. Inf. Comput. Sci. 1999, 39, 610–621. [Google Scholar]
- Lucic, B.; Amic, D.; Trinajstic, N. Nonlinear Multivariate Regression Outperforms Several Concisely Designed Neural Networks in QSPR Modeling. J. Chem. Inf. Comput. Sci. 2000, 40, 403–413. [Google Scholar] [CrossRef] [PubMed]
- Polanski, J.; Walczak, B. The comparative molecular surface analysis (CoMSA): a novel tool for molecular design. Comp. Chem. 2000, 24, 615–625. [Google Scholar]
- Polanski, J.; Gieleciak, R.; Bak, A. The comparative molecular surface analysis (CoMSA) – a nongrid 3D QSAR method by a coupled neural network and PLS system: Predicting pKa values of benzoic and alkanoic acids. J. Chem. Inf. Comput. Sci. 2002, 42, 184–191. [Google Scholar]
- Polanski, J.; Gieleciak, R. The comparative molecular surface analysis (CoMSA) with modified uninformative variable elimination-PLS (UVE-PLS) method: application to the steroids binding the aromatase enzym. J. Chem. Inf. Comput. Sci. 2003, 43, 656–666. [Google Scholar]
- Polanski, J.; Gieleciak, R.; Wyszomirski, M. Comparative molecular surface analysis (CoMSA) for modeling dye-fiber affinities of the azo and antraquinone dyes. J. Chem. Inf. Comput. Sci. 2003, 43, 1754–1762. [Google Scholar] [CrossRef]
- Polanski, J.; Gieleciak, R.; Wyszomirski, M. Mapping dye pharmacophores by the comparative molecular surface analysis (CoMSA): application to heterocyclic monoazo dyes. Dyes Pigm. 2004, 62, 63–78. [Google Scholar]
- Polanski, J.; Gasteiger, J.; Jarzembek, K. Self - Organizing neural networks for screening and development of novel artificial sweetener candidates. Combin. Chem. High Throughput Screen. 2000, 3, 481–495. [Google Scholar]
- Polanski, J.; Gieleciak, R. Comparative molecular surface analysis: a novel tool for drug design and molecular diversity studies. Mol. Diversity 2003, 7, 45–59. [Google Scholar] [CrossRef]
- Polanski, J. Self-organizing neural networks for pharmacofore mapping. Adv. Drug Deliv. Rev. 2003, 55, 1149–1162. [Google Scholar] [PubMed]
- Polanski, J.; Bak, A. Modeling steric and electronic effects in 3D and 4D-QSAR schemes: Predicting benzoic pKa values and steroid CBG binding affinities. J. Chem. Inf. Comput. Sci. 2003, 43, 2081–2092. [Google Scholar] [PubMed]
- Polanski, J.; Gieleciak, R.; Bak, A. Probability issues in molecular design: Predictive and modeling ability in 3D-QSAR schemes. Comb. Chem. High T. Scr. In press.
- Kohonen, T. Self-Organization and Associative Memory, 3rd Edition ed; Springer Verlag: Berlin, 1989. [Google Scholar]
- Zupan, J.; Gesteiger, J. Neural Networks in Chemistry and Drug Design, 2nd Edition ed; Wiley–VCH: Weinheim, 1999. [Google Scholar]
- Melssen, W.J.; Smits, J.R.M.; Buydens, L.M.C.; Kateman, G. Tutorial: Using artificial neural networks for solving chemical problems. Part II. Kohonen self-organising feature maps and Hopfield networks. Chemometer. Intell. Lab. Syst. 1994, 23, 267–291. [Google Scholar]
- Kohonen, T. The Self-Organizing Map (SOM), http://www.cis.hut.fi/projects/somtoolb.shtml
- Hopfinger, A.J.; Wang, S.; Tokarski, J.S.; Jin, B.; Albuquerque, M.; Madhav, P.J.; Duraiswami, C. Construction of 3D QSAR models using the 4D QSAR analysis formalism. J. Am. Chem. Soc. 1997, 119, 10509–10524. [Google Scholar]
- Polanski, J.; Gieleciak, R.; Magdziarz, T. The grid formalism for the comparative molecular surface analysis: application to the CoMFA benchmark steroids, azo dyes and HEPT derivatives. J. Chem. Inf. Comput. Sci. In press.
- Centner, V.; Massart, D.L.; de Noord, O.E.; de Jong, S.; Vandeginste, B.M.V.; Sterna, C. Elimination of uninformative variables for multivariate calibration. Anal. Chim. Acta. 1996, 330, 1–17. [Google Scholar]
- Pilizota, T.; Lucic, B.; Trinajstic, N. Use of variable selection in modeling the secondary structural content of proteins from their composition of amino acid residues. J. Chem. Inf. Comput. Sci. 2004, 44, 113–121. [Google Scholar] [CrossRef]
- Gasteiger, J. Match3D; KMAP for the information see: http://www2.ccc.uni-erlangen.de
- Sybyl 6.5. program, available from the Tripos Inc., St. Louis, MO, USA: http://www.tripos.com
- HyperChem 5.0; available from HyperCube Inc., Gainesville, FL, USA: http://www.hyper.com
- MATLAB 6.5; available from The Mathworks Inc., Natick, MA, USA, http://www.mathworks.com
© 2004 by MDPI (http://www.mdpi.org). Reproduction is permitted for noncommercial purposes.
Share and Cite
Polanski, J.; Bak, A.; Gieleciak, R.; Magdziarz, T. Self-organizing Neural Networks for Modeling Robust 3D and 4D QSAR: Application to Dihydrofolate Reductase Inhibitors. Molecules 2004, 9, 1148-1159. https://doi.org/10.3390/91201148
Polanski J, Bak A, Gieleciak R, Magdziarz T. Self-organizing Neural Networks for Modeling Robust 3D and 4D QSAR: Application to Dihydrofolate Reductase Inhibitors. Molecules. 2004; 9(12):1148-1159. https://doi.org/10.3390/91201148
Chicago/Turabian StylePolanski, Jaroslaw, Andrzej Bak, Rafal Gieleciak, and Tomasz Magdziarz. 2004. "Self-organizing Neural Networks for Modeling Robust 3D and 4D QSAR: Application to Dihydrofolate Reductase Inhibitors" Molecules 9, no. 12: 1148-1159. https://doi.org/10.3390/91201148