55
Modeling of Biomolecular Systems Interactions, Dynamics, and Allostery Poster Session I
1-POS
Board 1
Estimation of Protein Secondary Structure from FTIR Spectra Using Wavelet Analysis and
Neural Networks
Sherif Abbas
1,3
, Parvez Haris
2
, Feride Severcan
1
, Mete Severcan
1
.
1
Middle East Technical University, Ankara, Turkey,
2
De Montfort University, Leicester, United
Kingdom,
3
Ain Shams University, Cairo, Egypt.
In order to improve the accuracy to predict protein secondary structure from Fourier transform
infrared (FTIR) spectra, a novel method for extraction of Protein FTIR features using wavelet
transform analysis was developed. Recording of proteins FTIR spectra was performed in
aqueous solution using spectrum using BRUKER Vectra 200 FT-IR Spectrometer. The protein
spectra preprocessd using Matlab and OPUS 5.5 software. Amide I region (1700 - 1600 cm-1) is
utilized and its wavelet analysis has been done in order to do the secondary structure analysis of
proteins from FTIR spectra . The secondary stucture of this proteins has been calculated also
from thier x-ray crystal structure using PDBsum bioinformatics tool. This calculated structure
has been used as atarget for our Artificial Neural Network (ANN). These results indicate that the
methodology introduced is effective and estimation accuracies are in some cases better than
those previously reported in the literature.
