UWCScholar
This repository serves as a digital archive for the preservation of research outputs from the University of the Western Cape.
Communities in UWCScholar
Select a community to browse its collections.
Recent Submissions
The importance of time: modelling network intrusions with long short-term memory recurrent neural networks
(University of the Western Cape, 2024) Staudemeyer Ralf Colmar; Omlin Christian W
We claim that modelling network traffic as a time series with a supervised learning approach, using known genuine and malicious behaviour, improves intrusion detection. To substantiate this, we trained long short-term memory (LSTM) recurrent neural networks with the training data provided by the DARPA KDD Cup 99 challenge. After preprocessing all frature, to improve information gain. we applied a number of intuitive steps to extract salient features,which resulted in the creation of a number of minimal fcature sets that could be used for detecting attack classes. The preprocessed KDD Cup 99 data was then used to test the performance of five very common and well-known classifiers. Decision trees. naive Bayes, Bayesian networks, feed forward neural network and support vector machines. Our results show a performance comparable to the winning entries of the KDD Cup 99 challengc. Finally, we applied the LSTM recurrent neural network classifier to the preprocessed data using the minimal feature sets. Our results show that the LSTM classifier provides superior performance in comparison to other strong static classifiers t rained. This is due to the fact that, LSTM learns to look back in time and correlate constructive collection records. For the' first time ever, we have demonstrated the usefulness of LSTM networks to intrusion detection.
Screening for novel cold-adapted nitrile hydratases in Antarctic metagenornic libraries
(University of the Western Cape, 2024) Williams, Wesley Trevor; Cowan Don
Nitrile hydratases (NHase) catalyse the hydrolysis of nitriles to their corresponding amides, the first step in the nitrile degradation pathway. The second step is the conversion of the amide to a carboxylic acid, catalysed by an amidase. The NHases published to date have been isolated from mesophilic and thermophilic bacteria. The aim of this study was to identify a novel cold-adapted NHase from psychrophilic and psychrotrophic bacteria. The identification of a cold-adapted NHase is of interest for its potential use in chemical and pharmaceutical production. Enzymes isolated from psychrophilic bacteria typically have a high specific activ ity and low thermal stabi lity to cope with the inhibition of chemical reactions caused by low temperatures. However, culture based screening for these activities tend to reisolate already characterised isolates amiable to culturing. To access the majority of the bacterial biodiversity that the uncultured bacteria represent, the metagenomic approach was used to screen for nitrile degrading activity. Metage-nomic libraries constructed from soil taken from the McMurdo Dry Valleys in Antarctica were used for both
functional and sequence based screens of NHases. Functional screens that were trialled were based on pH change and oligotrophic selective growth.
Aspects of the interrelation between hypertension and insulin resistance: a preliminary study
(University of the Western Cape, 2024) Nwabuisi Osuafor, Godswill; Dietrich Daneel
It is well known that some genetic factors and dietary factors. such as excessive salt intake and excessive caloric intake (resulting in obesity) are risk factors for hypertension. Fifty percent of all hypertensive patients are also insulin resistant. Both hypertension and insulin resistance are again risk factors for other cardiovascular diseases such as atherosclerosis and heart fai lure. The nature of the association between hypertension and insulin resistance has not heen clearly elucidated. Spontaneously hypertensive rats are the ideal models to study the aspects of the relationships between hypertension and insulin resistance. Models of high-fat feeding induce obesity, hypertension and insulin resistance and are thus used extensively to study hypertension because these models closely mimic some of the renal and cardiovascular changes found in human hypertensive pati ents. The present study was initiated to evaluate if insulin resistance will develop within 6 weeks in a model of high- fat diet induced hypertension and if so. to determine whether captopril will affect the presence of insulin resistance. This model should in future be used to study vascular reactivity to phenylephrine (PHE), acetylcholinc (AC!-1) and sodium ni troprusside (S~P) in hypertensive animals in the absence or presence of insulin resistance and in normotensive insulin resistant animals.
Synthesis of sulphonated and transition metal oxide doped Polymeric nanocomposites for application in design of Supercapacitors
(University of the Western Cape, 2024) Njomo Njagi; Iwuoha Emmanuel I.
To meet a fast-growing market demand for next generation portable electronic devices with higher performance and increased device functionalities, efficient electrical energy devices with substantially higher energy, power densities and faster recharge times such as supercapacitors are needed. The overall aim of this thesis was to synthesize nanostmctured su lphonated polyaniline and transition metal single, binary and ternary mixed oxide doped nanocomposites with electro-conductive properties. These nanocomposites were anchored on activated graphitic carbon and used in design of asymmetric supercapacitors. Tan talum(IV)oxide, tantalum(IV)oxide-n ickel (II)oxide, tantalum(ll)oxide- manganese(IIT)oxide, tantalum(II)oxide-nickel(II)oxidemanganese( l l,IIl)oxide nanoparticles were synthesised using modified sol-gel methods. These were then dispersed, individually, in acidic media through sonication and
incorporated in-situ into the polymeric matrix during the oxidative chemical polymerization of aniline doped with poly(4-styrene sulphonic acid). These novel polymeric nanocomposites were characterised with FTIR, UV-visible, TEM, SEM, EDS, XRD to ascertain successful polymerization, doping, morphology and entrapment of the metal oxide nanoparticles. SECM approach curves and interrogation of CV revealed that these nanocomposites are conductive and electro-active.
Metagenomic bioprospecting: discovery and characterization of a novel esterase for biomass conversion
(University of the Western Cape, 2024) Ohlhoff Colin W.; Tuffin Marla
Due to an increasing human population and ever-expanding industrial economies, fossil fuel reserves are rapidly being depleted. In order to reduce our dependence on non-renewable energy resources much global interest has been directed towards the development and production of biofuels. Plant biomass has been identified as a possible feedstock for alternative fuel production and typically consists of cellulose, hemicellulose and lignin. Although these components can serve as valuable substrates in the bioconversion process, they are generally recalcitrant to enzymatic degradation. Current hydrolytic methods limit the depolymerization of the lignocellulose component of plant biomass, which in turn decreases the yield of potential fermentable sugars. As microorganisms represent an intriguing, underexploited resource for the discovery of novel biocatalysts, this study employed a metagenomic approach towards the identification of enzymes for
lignocellulose hydrolysis. A large-insert metagenomic fosmid library was constructed using total DNA extracted from thermophilic compost samples. The library comprised 150 000 clones, with an average insert size of 31 Kb, representing approximately 1300 prokaryotic genomes. Initial emphasis was placed on the assessment of the bacterial phylogenetic make-up of the metagenomic library and its validity for the mining of functional enzymes.