How can Python users be empowered with the robust simulation, compilation and scripting abilities of a non-proprietary object-oriented, equation based modeling language such as Modelica? The immediate objective of this thesis work is to develop an application programming interface for the OpenModelica modeling and simulation environment that would bridge the gap between the two agile programming languages Python and Modelica. The Python interface to OpenModelica OMPython, is both a tool and a functional library that allows Python users to realize the full capabilities of Open- Modelica’s scripting and simulation environment requiring minimal setup actions. OMPython is designed to combine both simulation and model building. Thus domain experts (people writing the models) and computational engineers (people writing the solver code) can work on one unified tool that is industrially viable for optimization of Modelica models, while offering a flexible platform for algorithm development and research.

Thesis (MSc)--Stellenbosch University, 2013. / ENGLISH ABSTRACT: Supply-demand analysis (SDA) is a tool that allows for the control, regulation and behaviour of metabolic pathways to be understood. In this framework, reactions are grouped into reaction blocks that represent the supply and demand of a metabolic product. The elasticities of these supply and demand blocks can be used to determine the degree of control either block has over the flux in the pathway and the degree of homoeostasis of the metabolic product that links the blocks. Rate characteristic plots, on which the rates of supply and demand blocks are plotted as functions of the concentration of the linking metabolite, represent a powerful visual tool in this framework. Generalised supply-demand analysis (GSDA) allows for the analysis of metabolic models of arbitrary size and complexity without prior knowledge of the regulatory structure of the pathway. This is achieved by performing SDA on each variable metabolite in a pathway instead of choosing a single linking metabolite. GSDA also provides other benefits over SDA as it allows for potential sites of regulation and regulatory metabolites to be identified. Additionally it allows for the identification and quantification of the relative contribution of di erent routes of regulation from an intermediate to a reaction block. Moiety-conserved cycles present a challenge in performing in silico SDA or GSDA, as the total concentration of a moiety must remain constant, thereby limiting the range of possible concentrations of the metabolites between which it cycles. The first goal of this thesis was to develop methods to perform GSDA on two-membered and interlinked moiety-conserved cycles. We showed that by expressing the members of a moiety-conserved cycle as a ratio, rather than individual metabolite concentrations, we can freely vary the ratio without breaking moiety conservation in a GSDA. Furthermore, we showed that by linking the concentrations of the members of two interlinked two-membered moiety-conserved cycles to a “linking metabolite”, we could vary the concentration of this metabolite, within constraints, without breaking moiety conservation. The Python Simulator for Cellular Systems (PySCeS) is a software package developed within our group that provides a variety of tools for the analysis of cellular systems. The RateChar module for PySCeS was previously developed as a tool to perform GSDA on kinetic models of metabolic pathways by automatically generating rate characteristic plots for each variable metabolite in a pathway. The plots generated by RateChar, however, were at times unclear when the models analysed were too complex. Additionally, invalid results where steady-states could not be reached were not filtered out, and therefore appeared together with valid results on the rate characteristic plots generated by RateChar. We therefore set out to improve upon RateChar by building plotting interface that produces clear and error-free rate characteristics. The resulting RCFigure class allows users to interactively change the composition of a rate characteristic plot and it includes automatic error checking. It also provides clearer rate characteristics with e ective use of colour. Using these tools two case studies were undertaken. In the first, GSDA was used to investigate the regulation of aspartate-derived amino acid synthesis in Arabidopsis thaliana. A central result was that the direct interaction of aspartate-semialdehyde (ASA), a metabolite at a branch point in the pathway, with the enzyme that produces it only accounts for 7% of the total response in the flux of supply. Instead, 89% of the observed flux response was due to ASA interacting with of the downstream enzymes for which it is a substrate. This result was unexpected as the ASA producing enzyme had a high elasticity towards ASA. In a second case study moiety-conserved cycles in a model of the pyruvate branches in lactic acid bacteria were linearised using the above mentioned method. This served to illustrate how multiple reaction blocks are connected by these conserved moieties. By performing GSDA on this model, we demonstrated that the interactions of these conserved moieties with the various reaction blocks in the pathway, led to non-monotonic behaviour of the rate characteristics of the supply and demand for the moiety ratios. An example of this is that flux would increase in response to an increase in product for certain ranges. This thesis illustrates the power of GSDA as an entry point in studying metabolic pathways, as it can potentially reveal properties of the regulation and behaviour of metabolic pathways that were not previously known, even if these pathways were subjected to previous analysis and a kinetic model is available. In general it also demonstrates how e ective analysis tools and metabolic models are vital for the study of metabolism. / AFRIKAANSE OPSOMMING: Vraag-en-aanbod analise (VAA) is ’n analisemetode wat mens in staat stel om die beheer, regulering en gedrag van metaboliese paaie beter te verstaan. In hierdie raamwerk word reaksies gegroepeer as reaksieblokke wat die aanbod (produksiestappe) en die aanvraag (verbruik-stappe) van ’n metaboliese produk verteenwoordig. Vanaf die elastisiteite van hierdie aanbod- en aanvraag-blokke kan die graad van beheer van elkeen van die blokke oor die fluksie, asook die graad van homeostase van die metaboliese koppelingsintermediaat, bereken word. Snelheidskenmerk-grafieke, waarop die snelhede van die vraag- en aanbod-blokke as funksies van die konsentrasie van die koppelingsmetaboliet uiteengesit word, verteenwoordig ’n kragtige visuele hulpmiddel in hierdie raamwerk. Veralgemeende vraag-aanbod analise (VVAA), die veralgemeende vorm van VAA, maak dit moontlikommetaboliese modelle van arbitrêre grootte en kompleksiteit te analiseer sonder enige vooraf-kennis van die regulatoriese struktuur van die paaie. Die prosedure is om VAA op elk van die veranderlike metaboliete in die pad uit te voer, eerder as om ’n enkele koppelingsmetaboliet te kies. VVAA het ook ander voordele bo VAA aangesien dit potensiële setels van regulering en regulatoriese metaboliete kan identifiseer. Daarbenewens kan dit die relatiewe bydrae van verskillende regulerings-roetes van vanaf ’n intermediaat na ’n reaksieblok identifiseer en hulle kwantifiseer. Groep-gekonserveerde siklusse bied ’n uitdaging vir in silico VAA of VVAA, aangesien die totale konsentrasie van die gekonserveerde groep konstant moet bly. Dit beperk die waardes van moontlike konsentrasies van die metaboliete wat die siklus uitmaak. Die eerste doelstelling van hierdie tesis was dus om metodes te ontwikkel waarmee VVAA op tweeledige en saamgebonde groep-gekonserveerde siklusse uitgevoer kan word. Deur die lede van groep-gekonserveerde siklusse eerder as verhoudings uit te druk in plaas van as individuele metabolietkonsentrasies, het ons gewys dat ons hierdie verhouding vrylik kan varieer sonder om die groep-konservering te breek in ’n VVAA. Ons het ook gewys dat die konsentrasies van die lede van ’n saamgebonde groep-gekonserveerde siklus gekoppel kan word aan ’n “koppelingsmetaboliet”, waarvan die konsentrasie dan binne perke gevarieer kan word sonder om die groep-konservering te breek. Die “Python Simulator for Cellular Systems” (PySCeS) is ’n programmatuur-pakket wat binne ons navorsingsgroep ontwikkel is met die doel om sellulêre sisteme numeries te analiseer. Die RateChar module vir PySCeS was reeds voor die aanvang van hierdie projek ontwikkel om VVAAop kinetiese modelle van metaboliese paaie uit te voer deur outomaties snelheidskenmerke vir elke veranderlikke metaboliet te genereer. Die grafieke wat deur RateChar gegenereer is, was egter soms onduidelik wanneer die modelle te groot of kompleks geraak het. Daarbenewens is ongeldige resultate, waar ’n bestendige toestand nie bereik kon word nie, nie uitgefiltreer nie, en het dus saam met geldige resultate op die snelheidskenmerke verskyn. Een van die doelstellings was dus om RateChar te verbeter deur ’n koppelvlak vir grafieke te ontwikkel wat duidelike en foutlose snelheidskenmerke kon produseer. Dit het gelei tot die RCFigure klas wat outomatiese foutopsporing uitvoer en gebruikers in staat stel om op ’n interaktiewe wyse die samestelling van ’n snelheidskenmerkgrafiek te verander. Dit bied ook duideliker snelheidskenmerke deur e ektief van kleur gebruik te maak. Met hierdie ontwikkelde gereedskap is twee gevallestudies onderneem. In die eerste is VVAA gebruik om die regulering van aspartaat-afgeleide aminosuursintese in Arabidopsis thaliana te bestudeer. Die belangrikste resultaat was dat die direkte interaksie van aspartaat-semialdehied (ASA), ’n metaboliet by ’n vertakkingspunt in die pad, met die ensiem wat dit produseer, slegs vir 7% van die totale respons in die aanbod-fluksie verantwoordelik was. Daarteen was 89% van die waargenome fluksierespons die gevolg van die interaksie van ASA met drie van die stroomafensieme, waarvoor dit ’n substraat is. Hierdie resultaat was onverwag aangesien die ensiem wat ASA produseer ’n hoë elastisiteit teenoor ASA toon. In ’n tweede gevallestudie is die groep-gekonserveerde siklusse in ’n model van die pirovaat-takke in melksuurbakterie-metabolisme gelineariseer deur gebruik te maak van die bo beskrewe metode. Dit illustreer hoe verskeie reaksieblokke verbind word deur hierdie gekonserveerde groepe. M.b.v. ’n VVAA van hierdie model het ons gedemonstreer dat die interaksies van die gekonserveerde groepe met die verskeie reaksieblokke in die pad kan lei tot nie-monotoniese gedrag van die snelheidskenmerke van die vraag- en aanbod-reaksies vir die verhouding van die gekonserveerde groep-komponente. ’n Voorbeeld hiervan is die onverwagte waarneming dat die fluksie toeneem met toenemende produk-konsentrasie oor sekere gebiede. Hierdie tesis illustreer die krag van VVAA as ’n beginpunt vir die studie van metaboliese paaie, aangesien dit onbekende regulatoriese eienskappe en gedragspatrone kan ontbloot, selfs al is die paaie vantevore m.b.v. kinetiese modelle geanaliseer. Oor die algemeen demonstreer dit die noodsaaklikheid van e ektiewe analisegereedskap en metaboliese modelle vir die bestudering van metabolisme. / National Research Foundation