Artificial Neural Network Modification of Simulation-Based Fitting: Application to a Protein-Lipid System

Simulation-based fitting has been applied to data analysis and parameter determination of complex experimental systems in many areas of chemistry and biophysics. However, this method is limited because of the time costs of the calculations. In this paper it is proposed to approximate and substitute a simulation model by an artificial neural network during the fitting procedure. Such a substitution significantly speeds up the parameter determination. This approach is tested on a model of fluorescence resonance energy transfer (FRET) within a system of site-directed fluorescence labeled M13 major coat protein mutants incorporated into a lipid bilayer. It is demonstrated that in our case the application of a trained artificial neural network for the substitution of the simulation model results in a significant gain in computing time by a factor of 5 x 10(4). Moreover, an artificial neural network produces a smooth approximation of the noisy results of a stochastic simulation.

[1]  George Cybenko,et al.  Approximation by superpositions of a sigmoidal function , 1992, Math. Control. Signals Syst..

[2]  J. A. Stegemann,et al.  A Glossary of Basic Neural Network Terminology for Regression Problems , 1999, Neural Computing & Applications.

[3]  Mikalai M. Yatskou,et al.  A study of energy transfer processes in zinc–porphyrin films using Monte Carlo simulation of fluorescence decay , 2001 .

[4]  R. Spruijt,et al.  Membrane-anchoring interactions of M13 major coat protein. , 2001, Biochemistry.

[5]  Gaudenz Danuser,et al.  FRET or no FRET: a quantitative comparison. , 2003, Biophysical journal.

[6]  Th. Förster.  Zwischenmolekulare Energiewanderung und Fluoreszenz , 1948 .

[7]  David Stopar,et al.  Structural characterization of bacteriophage M13 solubilization by amphiphiles. , 2002, Biochimica et biophysica acta.

[8]  Vladimir V. Apanasovich,et al.  Analysis of the Decay Kinetics of Fluorescence of Complex Molecular Systems Using the Monte Carlo Method , 2000 .

[9]  G. Pitsevich,et al.  The torsion potential function of hydrogen peroxide and disulfan molecules , 2000 .

[10]  J. Lakowicz.  Principles of fluorescence spectroscopy , 1983 .

[11]  Hans C. Gerritsen,et al.  Dynamic Monte Carlo simulations to model FRET and photobleaching in systems with multiple donor-acceptor interactions , 2002 .

[12]  George Cybenko,et al.  Approximation by superpositions of a sigmoidal function , 1989, Math. Control. Signals Syst..

[13]  Kurt Hornik,et al.  Multilayer feedforward networks are universal approximators , 1989, Neural Networks.

[14]  Vladimir V. Apanasovich,et al.  Nonisotropic Excitation Energy Transport in Organized Molecular Systems: Monte Carlo Simulation-Based Analysis of Fluorescence and Fluorescence Anisotropy Decay , 2001 .

[15]  David Stopar,et al.  Protein-lipid interactions of bacteriophage M13 major coat protein. , 2003, Biochimica et biophysica acta.

[16]  T.,et al.  Training Feedforward Networks with the Marquardt Algorithm , 2004 .

[17]  Philip D. Wasserman,et al.  Neural computing - theory and practice , 1989 .

[18]  Douglas B. Kell,et al.  Non-linear optimization of biochemical pathways: applications to metabolic engineering and parameter estimation , 1998, Bioinform..

[19]  R. Spruijt,et al.  Localization and rearrangement modulation of the N-terminal arm of the membrane-bound major coat protein of bacteriophage M13. , 2000, Biochimica et biophysica acta.

[20]  Heekuck Oh,et al.  Neural Networks for Pattern Recognition , 1993, Adv. Comput..

[21]  L. Stryer.  Fluorescence energy transfer as a spectroscopic ruler. , 1978, Annual review of biochemistry.

[22]  Igor V. Tetko,et al.  Neural network studies, 1. Comparison of overfitting and overtraining , 1995, J. Chem. Inf. Comput. Sci..