Search

Corporate governance plays a crucial role in increasing the access to external funding by firms, lowering cost of funding and boosting firm value, increasing operational performance, and decreasing financial crises risk. Moreover, sound corporate governance practices are essential to achieve and maintain public trust in banks in particular, where poor corporate governance can contribute to bank failures, which in turn poses significant macroeconomic consequences. Due to the important financial intermediation role of banks, the public have a high degree of sensitivity to difficulties arising from corporate governance deficiencies in banks. Consequently, the interrelation between corporate governance mechanisms and bank efficiency, valuation, and stability is still triggering more research, but until today, there is no consensus on the definite impact of the corporate governance frameworks on bank overall performance, and every study on this topic presents different results, which is attributed to the studied sample, covered period, and adopted empirical methodology. This thesis participates in this continuous debate by studying a sample formed of the largest banks operating in the MENA region over the period 2011-2018. The MENA banks represent an interesting case study because: a considerable number of them are state-owned, a considerable number of them are Islamic banks, they are characterised by considerable high ownership concentration, they have low gender diversity, and a considerable number of them adopt a CEO-chairman role duality. This thesis examines the impact of nine corporate governance variables (which fall under two main categories: ownership structure and board of directors' composition) on MENA banks' operational performance and efficiency (in chapter two), value (in chapter three), and risk and stability (in chapter four). Moreover, to detect if and how bank type (conventional or Islamic) shapes the relationship between corporate governance variables and the adopted dependent variables, where the sample under study is split between conventional and Islamic banks in the econometric estimations. Indeed, the obtained results reveal that the explanatory variables influence the two categories of banks differently. These results allow drawing several policy recommendations for MENA banks and for regulatory authorities regarding the optimal corporate governance structure that maximises operational performance, value, and stability.

The effect of hydrophobicity on susceptibility of phosphatidylcholine derivatives to PLA2 attack was studied by synthesising two series of compounds, the saturated symmetrically distributed derivatives and compounds with one long acyl-chain in the first position and a selected short-chain in the second position. The first series were used to investigate the effect of substrate morphology over the range from free monomer substrates through micelle to bilayer forms, the second series were used to examine the region of the micelle/bilayer transition at higher resolution. The symmetric substrates showed the expected properties with susceptibility increasing with hydrophobicity in the free monomer series and a sharp transition seen at the CMC for the shorter chain substrates. No such change was seen for longer chain substrates with any of these enzymes and the concept of a hydrophobic anchor increasing the rate of attack on condensed rather than monomeric substrates was questioned. These compounds were used to study the action of enzyme activated by treatment with oleoyl-imidazolide. The biggest activation factors were seen for asymmetric bilayer-forming substrates in propanolic solution, but the physico-chemical form of the substrates under these conditions was not established. Fatty acyl activation was found to alter the response to calcium activation and suggested a two-calcium site model. Studies of calcium activation showed very strong evidence for two kinetically important sites. Metal ion inhibition showed that barium and large cations were competitive inhibitors for calcium, but zinc and cadmium were not and appeared to inhibit a component of activation only found at high calcium activation. This lead to the proposal that zinc and barium bind to the enzyme at different calcium binding sites. Hydrolysis curves were shown to vary in shape depending on calcium concentration and the anomalous shape was associated with the presence of calcium at a single binding site. Addition of zinc removed the anomalous shape and without giving further enzyme activation whilst addition of calcium changes the curve shape and was activating. Some evidence was presented to suggest that the presence of a metal ion in the second (zinc) site was important for modulating the activation of the enzyme with surfaces. The use of the conductimetric assay was extended to the purification and characterisation of PLD enzymes acting as both transferases and hydrolases The transferase activity was used to generate phosphatidyl alcohols from DiC8PC and the methanol derivative was shown to be better substrate than the PC equivalent but to posess chartacteristics of the monomeric form. Polar phosphatidic acid compounds which should have high CMC values sould enable the relationship between structure, morphology and susceptibility to PLA2 to be examined in detail.

Abstract: Calculating and modeling a variogram from sparsely sampled data can be complex and time-consuming due to the need for expertise in selecting the appropriate parameters and fitting functions to the experimental variogram. To assist with variogram modeling, a novel approach based on Convolutional Neural Networks (CNNs) are implemented in this research to automatically calculate and model variograms from sparse data. CNNs are known for effectively processing grid-like data such as images and spatial maps making them suitable for geostatistical applications. To train the CNN model, a large set of Sequential Gaussian Simulation (SGS) realizations, each with different variogram model parameters, is resampled to generate the training data. Training a CNN directly on sampled data was not effective so each dataset is modeled using inverse distance; other interpolation techniques such as kriging and nearest neighbors were also tested, but inverse distance was found to be the most efficient for training data, the minimal number of input parameters for inverse distance is attractive. The approach consists of the implementation of three CNNs: CNN-A predicts the major direction of continuity, CNN-1 estimates key variogram model parameters, and CNN-2 predicts the experimental variogram values at pre-set lag distances. Considering these CNNs, two workflows are implemented: (1) train CNN-1 to directly predict variogram model parameters (range, anisotropy ratio, nugget effect) followed by the inference of an automated model based on the predicted parameters (2) train a second CNN-2 to predict the experimental variogram values at specified lag distances which are easily autofitted. Mean squared error (MSE) is used as the loss function in the training process. Once trained, the CNNs can predict the azimuth of the major direction, key variogram model parameters, and the experimental variogram of a given dataset. When applying CNN to datasets, data augmentation (rotation) is performed to obtain a set of variogram predictions that improve prediction robustness. Hyperparameter tuning and sensitivity analysis are performed to determine optimal parameters for the CNN by choosing configurations that minimize MSE and avoid overfitting. The testing and validation of variogram prediction are conducted on various datasets, including the Walker Lake, and a data validation consisting of 110 geological images transformed to grayscale values. The results indicate that the prediction of variogram model parameters (CNN-1) is limited by a fixed number of variogram structures and the second workflow (CNN-2) involving predicting values at pre-set lag distances performs better without limiting the number of variogram structures. CNN-2 results in the prediction of an experimental variogram that can be easily fit with automated variogram programs achieving an R-squared value of 0.95 for the validation dataset. The workflows are fully automated and applicable to sparse or dense data but are currently limited to 2D datasets and the output consists of normal score variograms.

This work addresses current problems of RNA sequence alignment and describes different tools for solving them. RNA molecules form basepairs that fold the molecule into its secondary and tertiary structures. These structures are more conserved in evolution than primary sequence because they directly affect the function of the molecule. Thus, sequence alignment of RNA molecules, unlike that of other biological molecules, must proceed by aligning homologous pairs of positions to each other. The state of the art methods used today for aligning RNA are based on Stochastic Context Free Grammars (SCFG). These methods are able to characterize and thus align nested RNA basepairs, but are incapable of dealing with crossing basepair patterns. In addition, the current application of this algorithm ignores 3D structure and thus deals best with only one type of basepairs. Although this type (the cis Watson-Crick/Watson-Crick or cWW type) is the most common in RNA, there are at least eleven other families of basepairs that account for about one third of RNA interactions. Each of these families has its own structural dimension, and therefore its own patterns of accepted isosteric substitutions in sequence alignments. Here, an RNA alignment analysis and evaluation tool that takes into consideration 3D structure with all types of basepairs is described. This tool is used to structurally evaluate alignments and locate errors in them. A discussion and classification of tertiary interactions of the G/U wobble basepair is then presented. Novel conserved interactions are discovered, and their sequence signatures are used to further enhance sequence alignments. Finally, a better SCFG approach for automatic RNA alignment is suggested and tested. This approach takes into consideration the 3D structure of all families of basepairs. It is also coupled with another theory, Markov Random Field (MRF), to align areas where crossing basepairs occur.