Browsing by Author "Louw, Carla"

Now showing 1 - 2 of 2
  • Thumbnail Image
    Item
    Comparison of progress curve analysis and initial rate kinetics for the estimation of identifiable enzyme kinetic parameters : a case study using two glycolytic enzymes from Sulfolobus solfataricus
    (Stellenbosch : Stellenbosch University, 2016-12) Louw, Carla; Snoep, Jacky L.; Eicher, Johann J.; Stellenbosch University. Faculty of Science. Department of Biochemistry.
    ENGLISH ABSTRACT: Obtaining accurate parameter value estimations is imperative to studying enzyme kinetics. It is also important that these parameter values be identifiable. The nonidentifiability of parameter estimations is usually indicative of parameter correlation, but may also be caused by other factors. These factors include the quality and quantity of the data used for parameter characterisation, the method of data analysis used as well as the model or rate equation to which the data is fitted. The non-identifiability of parameter estimations can either be structural or practical non-identifiability. Structural non-identifiability is caused by the model structure and is an indication of parameter correlations while practical non-identifiability is an indication of qualitatively or quantitatively insufficient data. The choice in kinetic assay and the type of data that will be collected is often dictated by the ligands, the enzyme itself or the availability of equipment. However, in the absence of these limitations, either progress curve or initial rate analysis may be selected. There are advantages and limitations to using either of these methods. The aim of this study is to try and determine if either of these two methods returns a greater number of identifiable parameter estimations for specific enzyme kinetic attributes. Experimental progress curve and initial rate data are collected for two different enzymes, GAPN and PGI, of Sulfolobus solfataricus. The GAPN and PGI enzymes represent different enzyme kinetic characteristics. Parameter value estimations are then obtained from the data fitting and an identifiability analysis for all possible combinations of the parameters are completed. When considering all the combinations of parameter fits it is possible to see trends in the combinations that return identifiable or non-identifiable parameters. With the identifiability analysis approach used in this study it is possible to determine if the non-identifiability of a parameter is structural or practical. Thereafter it is possible to speculate if the non-identifiability of the parameter estimations are due to parameter correlation or other factors and which method is superior for the analysis of certain enzyme kinetic attributes.
  • Thumbnail Image
    Item
    Computational modelling of steroid hormone biosynthesis and metabolism
    (Stellenbosch : Stellenbosch University, 2020-04) Louw, Carla; Snoep, Jacob Leendert; Van Niekerk, David Douglas; Stellenbosch University. Faculty of Science. Dept. of Biochemistry.
    ENGLISH ABSTRACT: This study describes the use of computational modelling and statistical techniques to address three topics in the field of steroid hormone research. The first is the hypocortisolism seen in the South African Angora goat. In a comparative analysis the construction, parameterisation and validation of a model describing the D4 and D5 pathways in both the ovine and Angora goat species are completed. With these models the issue of identifying a possible treatment target is addressed. The second topic is the steroidogenic activity in the human liver. This includes the construction, parameterisation and validation of a model describing the relative conversion of classic androgens and 11-oxygenated androgens to their respective 5a or 5b reduction products in the human liver. The model indicates that under physiological steady state conditions the 5b reduction of both the classic androgens and the 11-oxygenated androgens are the preferred reaction. The third and final topic discussed in this study is the steroidogenic activity in prostate cancer C42B cells. The construction, parameterisation and validation of a model describing the steroid hormone biosynthesis in the C42B castration resistant prostate cancer cell line is included. Three possible treatment targets for castration resistant prostate cancer in C42B cells are identified. This study also describes the development of an add-on Mathematica package, IdentifiabilityAnalysis, which simplifies the process of model parameterisation and identifiability analysis. This package is used throughout this study.