Computational studies of integrin inhibitors

Alarfaji, Saleh (2018) Computational studies of integrin inhibitors. PhD thesis, University of Nottingham.

[img] PDF (This version contains corrections, which need to be approved by examiner. Also, it contains commercial sensitive data with embargo of 5 years) (Thesis - as examined) - Repository staff only until 1 September 2023. Subsequently available to Repository staff only - Requires a PDF viewer such as GSview, Xpdf or Adobe Acrobat Reader
Download (18MB)

Abstract

Compounds containing amides play a key role in the pharmaceutical industry, with one area of interest, amongst many, being the development of new antifibrotic therapies. These compounds can bind to different integrins, which are essential activity regulators for transforming growth factor beta 1 (TGF-β1), and biological activity can assessed using a cell adhesion assay. Two- and three-dimensional quantitative structure-activity relationships (2D/3D-QSARs) were used to model the correlation between the physicochemical properties of some amides and their biological activity to predict the activities of new molecules. An autocorrelation based method,topological maximum cross correlation (TMACC), was employed to build our 2DQSAR models. We have generated models across four sets of data, ranging in size

from 25 to 46 molecules on a number of integrins subtypes. The results were crossvalidated using a leave-one-out (LOO) approach. Using partial-least-squares regression, a TMACC model with good predictive ability was generated based on training set of 25 compounds and showed satisfactory statistical results (q2 = 0.57, r2 = 0.82). Other TMACC models for 40 and 46 molecules were generated and showed similar but slightly higher predictivity: (q2 = 0.71, r2 = 0.93), and (q2 = 0.49, r2 = 0.70), respectively. 3D-QSAR models were also established for the same data sets using comparative molecular field analysis (CoMFA) and comparative molecular similarity indices analysis (CoMSIA). The CoMFA and CoMSIA models were not sensitive to changes in the orientation of the amide structure. The TMACC QSARs showed better predictive ability than the 3D-QSARs. We also have employed a quantum chemical investigation into different datasets of benzamides in order to understand the relationship between the physicochemical properties of all datasets with their activity. Potential energy surface scans were calculated for each molecule using density functional theory (DFT). Several quantum descriptors were calculated and used to build our QSAR models. QSAR results were validated using the LOO and Y-randomisation methods. Statics were also employed to determine all possible correlation between the quantum descriptors and biological activity.

In a distinct study, the electronic transitions, in the aromatic side chain of the amino acid phenylalanine toluene was investigated. The toluene molecule was used as a model system for biological chromophores. The photophysics of toluene was studied to understand the photophysics of large biomolecules such as proteins better. The ground (S0) and excited (S1) states in the gas phase have been calculated using a multi-configurational CASSCF method. We also employ the state-averaged CASSCF approach to estimate the geometries and harmonic vibrational frequencies for the both S0 and S1 states too. The ππ∗ transition is calculated to be at 4.69 eV, and the corresponding experimental energy is 4.65 eV. The experimental ππ∗ transition dipole moment is 0.37 D. If four unoccupied a´ orbitals are included in the active space, the calculated ππ∗ transition dipole moment is 0.30 D. If the number of unoccupied a´ orbitals is decreased to six orbitals, then the calculated ππ∗ transition moment drops to 0.28 D.

In the future, further improvements of TMACC and 3D-QSAR are needed. For a given data set, the TMACC descriptor construction should be enlarged to include other essential and new atomic properties. Moreover, the output of TMACC and 3D-QSAR would be more beneficial if we could automate feature detection for any large data.

Item Type: Thesis (University of Nottingham only) (PhD)
Supervisors: Hirst, Jonathan D.
Faculties/Schools: UK Campuses > Faculty of Science > School of Chemistry
Item ID: 55472
Depositing User: Alarfaji, Saleh
Date Deposited: 01 Oct 2021 14:53
Last Modified: 01 Oct 2021 14:54
URI: http://eprints.nottingham.ac.uk/id/eprint/55472

Actions (Archive Staff Only)

Edit View Edit View