Araree Jirapornanan. Development of Enzyme Kinetic Module Tool for Enzymatic Reaction Rate Prediction. Master's Degree(Bioinformatics). King Mongkut's University of Technology Thonburi. Library. : King Mongkut's University of Technology Thonburi, 2004.
Development of Enzyme Kinetic Module Tool for Enzymatic Reaction Rate Prediction
Organization :
King Mongkut's University of Technology Thonburi. School of Bioresources and Technology and School of Information Technology. Bioinformatics
Abstract:
Hybrid model of glycolysis pathway of Saccharomyces cerevisiae was constructed by
inserting neural network into mass balance equations. The performance of hybrid model
was tested by comparing results which are concentrations of metabolites to the results of
the original model derived from the work of J.L Galazzo and J.E. Bailey (1990). Two
networks representing enzyme reaction rate simulators for two enzymes,
Glyceraldehyde-3-Phosphate Dehydrogenase (GAPDH) and Phosphofhctokinase
(PFK), were constructed. The proposed networks were designed to have five nodes in
input layer for five variables; glucose (Glc), Glucose 6-phosphate (G6P), fructose-1,6-
diphosphate (FDP), phosphoenol pyruvate (PEP), and Adenosine triphosphate (ATP);
and one node in output layer for reaction rate of an enzyme. The concentrations of
metabolites and enzyme reaction rates of GAPDH and PFK as function of time obtained
fiom the original model were used as training data for two networks, respectively. The
trained networks were used to construct three hybrid models; first with the network of
enzyme GAPDH, second with the network of enzyme PFK, and third with two networks
of both enzymes. Mean square errors of results between the original and hybrid models
were obtained in the range between 3.36~o1V 4to 7.561~0 -lo. The sensitivity analysis
was performed by increasing and decreasing 10% and 20% of the initial concentrations
of metabolites and mean square errors were obtained in the range between 1.69~10t-o~
4.32~10-I. This work has provided the evidence for the potential of using hybrid
approach for biological systems modeling such as cellular dynamics which are smaller
scale.
Abstract:
VPI major capsid protein of polyomavirus has recently attracted much attention as
pharinaceutical carriers for DNA vaccination because of its ability to form the virus-like
particle without VP2/3 minor capsid proteins. To use VPl protein as particle-based
vaccine, understanding of the capsid's properties at molecular level is required. One
remaining problem is the assembly mechanism to form the capsid. This project aimed to
find out the residues of VPl capsid protein that are important to assembly function of
polyomavirus. The residues should be conserved throughout the evolution and show
high evolutionary conservation score. Mutation on the residues should result in
degradation of capsid formation. The conservation score for each arnino acid of lsva
protein, a VPl protein of simian virus 40; ldzl protein, Ll protein of human
papillomavirus 16; and lvps protein, polyomavirus VPI pentalner complexed; were
calculated based on amino acid frequency and replacements through evolutionary period
by using PAM250 scoring matrix. Conserved residues were analyzed whether they are
related to assembly function by checking the results of point mutations done on VPI
protein before. The conservation scores calculated based on both concepts are required
to define the highly conserved residues, some of which are consistent to the residues
informed by literature. The conservation scores calculated here is an alternative to
suggest positions to experimentalists for sequence modifications in order to study the
assembly process. In the long tenn, this study would lead to improvement of virus-like
particle transport properties.