Loading...

Linear and Non-Linear Modeling of Electrophoretic Mobility of Peptides

Darvizeh, Fatemeh | 2009

572 Viewed
  1. Type of Document: M.Sc. Thesis
  2. Language: Farsi
  3. Document No: 39563 (03)
  4. University: Sharif University of Technology
  5. Department: Chemistry
  6. Advisor(s): Jalali Heravi, Mehdi
  7. Abstract:
  8. Regarding the importance of biological systems in daily life and the complexity of these systems, this project is concerned with this problem and especially with applications of chemometrics in proteomics. In this respect, specific importance of peptides has been taken into account in the process of construction of especial and necessary proteins for human body. Due to the risks involved in some experimental investigations, it is quite preferable to utilize modeling approaches using different sets of data. Achieving a number of specific descriptors, a powerful can be established. This model could be quiet comprehensive for the prediction of the electrophoretic mobility of peptides. This model not only is able to predict electrophoretic mobility of peptides but also can describe some thermodynamic and electrostatic properties of peptides. This is a great achievement for major biological systems. The purpose of this thesis is to construct a robust model to predict the electrophoretic mobility of peptides .Two nonlinear ANN and SVM models are proposed in this work. Inputs of these models were obtained from a step-wise MLR model. Comparing the nonlinear ANN and SVM models with linear MLR model clearly indicates that nonlinear models are quite superior in prediction of electrophoretic motilities. The results obtained from two methods; ANN and SVM are as follow: R2 =0.96 and R2 =0.92 respectively. In spite of the result of ANN method is greater but the SVM method is more reapetable. Some other advantages of this research are achieving a series of constitutional, electrostatic and thermodynamic descriptors. the results obtained from the nonlinear models help in deduction of hidden concepts in a large number of data. This is a new and valuable achievment in the field of chemometrics.

  9. Keywords:
  10. Nonlinear Modeling ; Chemometrics Method ; Electrophoretic ; Peptides

 Digital Object List

 Bookmark

No TOC