Loading...

Development of a Model for Prediction of Inhibitors of HIV1 Virus

Hakimi, Fatemeh | 2010

858 Viewed
  1. Type of Document: M.Sc. Thesis
  2. Language: Farsi
  3. Document No: 41434 (03)
  4. University: Sharif University of Technology
  5. Department: Chemistry
  6. Advisor(s): Jalali Heravi, Mehdi
  7. Abstract:
  8. The main aim of this study is developing a robust QSAR model for describing and predicting the inhibitory activities of O-(2-phthalimidoethyl)-N-substituted thiocarbamates derivatives as novel HIV-1 non-nucleoside reverse transcriptase (HIV-1 NNRTIs) inhibitors. These drugs change the active site of the reverse transcriptase enzyme, and finally halter the HIV reproduction cycle. As the first step of this study, a multiple linear regression (MLR) model was built but it has no satisfied prediction ability. As a next step, the nonlinear correlation of the molecular descriptors and activities has been investigated by using artificial neural networks (ANN). In this section the effects of variable selection methods for construction of ANN models were examined. Three variable selection methods of stepwise MLR, genetic algorithm-partial least squares (GA-PLS) and genetic algorithm-artificial neural network (GA-ANN) were used and their results were compared. R2 values for leave-one-out cross validation (LOO-CV) procedure revealed that GA-ANN technique is the best algorithm for modeling of the activities. The advantages of GA-ANN over MLR-ANN and GA-PLS-ANN are its ability for variable selection and modeling, simultaneously. The GA-ANN descriptors revealed that the amounts of twigged and spatial shape of molecules are the most important parameters for changing the structure of the enzyme and locking its activity
  9. Keywords:
  10. Quantitative Structure-Activity Relationship (QSAR)Model ; Chemometrics Method ; Artificial Neural Network ; Genetic Algorithm ; Acquired Immune Deficiency Syndrome (AIDS)

 Digital Object List