Akademik Veri Yönetim Sistemi
Tezin Türü: Doktora
Tezin Yürütüldüğü Kurum: Orta Doğu Teknik Üniversitesi, Türkiye
Tezin Onay Tarihi: 2016
Tezin Dili: İngilizce
Öğrenci: Sherif Abbas Mousa Abbas
Danışman: FERİDE SEVERCAN
Özet:Proteins play very important roles in cells regulation and structure. Understanding of proteins structures greatly help in understanding of the mechanism of action of these proteins. The optical spectroscopic techniques such as Fourier transform infrared (FTIR) and circular dichroism (CD) spectroscopy can be used to study proteins in its native environment without complicated sample preparations that is required for a high resolution technique such as x-ray. In order to calculate the protein secondary structure from amide I band in FTIR spectra, different methods can be used such as curve fitting and deconvolution. However, these techniques have some disadvantages due to the noise and dependence on the operator. In this study, a protein FTIR dataset of known proteins structures was produced using FTIR transmission mode. This database was used as a training set for an artificial neural networks (ANNs). Because of the limited number of our proteins in the dataset (35 proteins), a leave-one-out approach for training and testing our neural networks was performed. To achieve generalized ANNs in a limited number proteins dataset, discrete wavelet transform (DWT) was successfully used as a data reduction technique for amide I spectra. The results of ANNs predictions showed 96.88%, 93.92% and 95.98% success in β-sheets, α-helix and other structures respectively. In the second part of this thesis, Human Apo- and Holo-transferren structure and their thermal denaturation behavior in dilute and artificial crowded environment was studied using FTIR spectroscopy. Dextran 70 and Ficoll 70 as a “molecular crowder” did not have a major effect on the secondary structure of transferrin as deduced from the analysis of the amide I band. However, it does alter the tertiary structure since significant differences in hydrogen-deuterium exchange was seen by monitoring the intensity of the residual amide II band as a function of time. The study of transferrin thermal denaturation using 2D-IR showed two different aggregated secondary structures patterns in dilute and in an artificial crowded environment. Finally, the proteins secondary structure of human pleural fluid accumulated due to malignant pleural mesothelioma (MPM), lung cancer (LC) and benign transudate (BT) was studied using attenuated total reflectance FTIR spectroscopy. Wavelet analysis was performed to extract the amide I spectral features. The extracted features were used as an input for the previously trained artificial neural network in the first part of this thesis. The ANNs results indicated significant differences in protein content of BT, LC and MPM pleural fluid samples. The chemometric results of the plural fluid proteins secondary structure lead to an accurate, cost effective method for the diagnosis of MPM from lung cancer and benign transudate with 88% sensitivity and 100% specificity.