Computational and analytical methods for constructing a multilevel model for human glucose metabolism
| dc.contributor.advisor | Snoep, Jacob Leendert | en_ZA |
| dc.contributor.advisor | Van Niekerk, David Douglas | en_ZA |
| dc.contributor.advisor | Cang, Hui | en_ZA |
| dc.contributor.author | Green, Kathleen Alice | en_ZA |
| dc.contributor.other | Stellenbosch University. Faculty of Science. Dept. of Mathematical Sciences. | en_ZA |
| dc.date.accessioned | 2022-03-11T09:28:18Z | |
| dc.date.accessioned | 2022-04-29T12:55:09Z | |
| dc.date.available | 2022-09-11T03:00:09Z | |
| dc.date.issued | 2022-03 | |
| dc.description | Thesis (PhD)--Stellenbosch University, 2022. | en_ZA |
| dc.description.abstract | ENGLISH ABSTRACT: Glucose metabolism is carefully regulated in humans to ensure that homeosta- sis is maintained. Disruptions in the multiple processes involved, or the inabil- ity to sustain adequate glucose concentrations, can cause various metabolic complications that can become life threatening. These complications can present as a result of diseases such as Type 2 diabetes (hyperglycaemia), or severe malaria (hypoglycaemia). In the context of malaria, two key metabolic indicators for poor chance of survival are hypoglycaemia (low plasma glu- cose concentrations) and lactic acidosis (high plasma lactate concentrations). Currently, it is understood that these conditions are the result of various clin- ical complications, and the extent to which the malaria parasite Plasmodium falciparum contributes to them is unknown. This contribution could be a consequence of the accelerated glycolytic flux, brought about by the parasite increasing the glucose demand and lactate production, once it has invaded the host’s red blood cells, a hypothesis that is tested in this thesis using a mathematical modelling approach. We used a new approach to building a whole body glucose metabolism model that is well-grounded in a large number of clinical studies following a thorough literature review to obtain clinical data on glucose metabolism. This model is parametrised using data from 49 different studies, and 74 figures that have been successfully reproduced between different softwares. The model construction is performed using a specialised package for model merging called Hierarchical Model Composition [1]. This model consists of several different organs that contribute to glucose metabolism in humans with a specific compartment that was incorporated to describe red blood cell metabolism. In addition to the reference model built for glucose metabolism in a healthy individual, we extend the model to represent malaria patients by explicitly modelling parasitaemia via the inclusion of a detailed mathematical model for Plasmodium falciparum into the red blood cell compartment. The multilevel model for malaria reveals that a 13% parasite burden leads to hypoglycaemia, but lactic acidosis as is observed in malaria patients, is not induced. Patient data and sensitivity analysis is used for initial model validations and identification of potential treatment targets in the parasite’s glycolytic pathway. The multilevel model is large (303 variables) which makes it difficult to anal- yse. Therefore we developed a flexible model reduction technique that can aid in the simplification of the multilevel model through selection of the relevant enzyme mechanisms, while retaining the whole body descriptions on the higher level. This reduction method applies a combination of structural and kinetic modification to the original model, and was tested on different modelling struc- tures and kinetics occurring in biochemical pathways. Thereafter, the method is extended to biological applications which show how multiple model simpli- fications for different inhibitor titration studies can be investigated starting from a single model description and performing various selections of reactions or species. During model merging we encountered logistical challenges such as unit con- version and the use of unique identifiers that are generic for merging different modules into a single model. Our solution was to use automated approaches as much as possible as developed in the Systems Biology community, and to code additional solutions for our automated workflow. This work highlights the benefits of utilising automated approaches, as well as combining differ- ent computational and analytical techniques from different disciplines, during model construction, validation and analysis. By making these models, all datasets, and simulation experiment descriptions available on JWS Online [2], FAIRDOMHub [3], and PK database [4], we envisage that future improve- ments and extensions can be implemented in a systematic way owing to the modular structure of the model, and the transparency and reproducibility of the construction process. | en_ZA |
| dc.description.abstract | AFRIKAANSE OPSOMMING: Glukosemetabolisme word noukeurig by mense gereguleer om te verseker dat homeostase gehandhaaf word. Ontwrigtings in die veelvuldige prosesse be- trokke, of die onvermoë om voldoende glukosekonsentrasies te volhou, kan verskeie metaboliese komplikasies veroorsaak wat lewensgevaarlik kan word. Hierdie komplikasies kan voorkom as gevolg van siektes soos tipe 2-diabetes (hiperglukemie), of ernstige malaria (hipoglukemie). In die konteks van malaria is twee sleutel metaboliese aanwysers vir ’n swak kans op oorlewing hipoglukemie (lae plasmaglukosekonsentrasies) en hiperlaktatemie (hoë plasmalaktaatkon- sentrasies). Tans word hierdie kondisies aan verskeie kliniese komplikasies toegeskryf, en die mate waarin die malariaparasiet Plasmodium falciparum daartoe bydra, is onbekend. Hierdie bydrae kan ’n gevolg wees van die ver- snelde glikolitiese fluksie wat veroorsaak word deur die parasiet wat die glukoseaan- vraag en laktaatproduksie verhoog sodra dit die gasheer se rooibloedselle bin- negedring het, ’n hipotese wat in hierdie tesis getoets word deur ’n wiskundige modelleringsbenadering te gebruik. Ons het ’n nuwe benadering gebruik om ’n hele-liggaam glukosemetabolisme model te bou wat goed gegrond is in ’n groot aantal kliniese studies na ’n om- vattende literatuuroorsig om kliniese data oor glukosemetabolisme te verkry. Hierdie model is geparametriseer deur gebruik te maak van die data van 49 verskillende studies, en 74 figure wat suksesvol tussen verskillende sagteware gereproduseer is. Die modelkonstruksie word uitgevoer deur gebruik te maak van ’n gespesialiseerde pakket vir modelsamesmelting genaamd Hierarchical Model Composition [1]. Hierdie model bestaan uit verskeie organe wat by- dra tot glukosemetabolisme by mense, met ’n spesifieke kompartement wat ingesluit is om rooibloedselmetabolisme to beskryf. Benewens die verwys- ingsmodel wat vir glukosemetabolisme in ’n gesonde individu gebou is, brei ons die model uit om malariapasiënte voor te stel deur parasitemie eksplisiet te modelleer met ’n gedetailleerde wiskundige model vir Plasmodium falci- parum in die rooibloedselkompartement. Hierdie multivlakmodel vir malaria toon dat ’n 13% parasietlas lei tot hipoglukemie, maar hiperlaktatemie, soos waargeneem by malariapasiënte, word nie geïnduseer nie. Pasiëntdata en sen- sitiwiteitsanalise word gebruik vir aanvanklike modelvalidasies en identifikasie van potensiële behandelingsteikens in die parasiet se glikolitiese pad. Die multivlakmodel is groot (303 veranderlikes) wat dit moeilik maak om te analiseer. Daarom het ons ’n buigsame modelreduksietegniek ontwikkel wat kan help met die vereenvoudiging van die multivlakmodel deur die keuse van die relevante ensiemmeganismes, terwyl die hele-liggaamsbeskrywings op die hoër vlak behou word. Hierdie reduksiemetode pas ’n kombinasie van strukturele en kinetiese veranderinge op die oorspronklike model toe, en is getoets op verskillende modelleringstrukture en kinetika wat in biochemiese padweë voorkom. Daarna word die metode uitgebrei na biologiese toepass- ings wat wys hoe veelvuldige modelvereenvoudigings vir verskillende inhibitor- titrasiestudies ondersoek kan word deur te begin met ’n enkele modelbeskry- wing en verskeie keuses van reaksies of spesies uit te oefen. Tydens modelsamesmelting het ons logistieke uitdagings teëgekom wat gener- ies is vir die samevoeging van verskillende modules in ’n enkele model, soos eenheidsomskakeling en die gebruik van unieke identifiseerders. Ons oplossing was om geoutomatiseerde benaderings, soos ontwikkel in die Sisteembiologie- gemeenskap, sover moontlik te gebruik en om bykomende oplossings vir ons geoutomatiseerde werksvloei te kodeer. Hierdie werk beklemtoon die voordele van die gebruik van geoutomatiseerde benaderings, sowel as die kombinasie van verskillende berekenings- en analitiese tegnieke uit verskillende dissiplines, tydens modelkonstruksie, validering en analise. Deur hierdie modelle, alle datastelle en simulasie-eksperimentbeskrywings beskikbaar te stel op JWS On- line [2], FAIRDOMHub [3] en PK-database [4], voorsien ons dat toekomstige verbeterings en uitbreidings geïmplementeer kan word in ’n sistematiese wyse te danke aan die modulêre struktuur van die model, en die deursigtigheid en reproduseerbaarheid van die konstruksieproses. | af_ZA |
| dc.description.version | Doctoral | en_ZA |
| dc.embargo.terms | 2022-09-11 | |
| dc.format.extent | xviii, 233 pages : illustrations (some colour), maps | en_ZA |
| dc.identifier.uri | http://hdl.handle.net/10019.1/125127 | |
| dc.language.iso | en_ZA | en_ZA |
| dc.publisher | Stellenbosch : Stellenbosch University | en_ZA |
| dc.rights.holder | Stellenbosch University | en_ZA |
| dc.subject.lcsh | Glucose -- Metabolism -- Mathematical models | en_ZA |
| dc.subject.lcsh | Parasitic diseases -- Mathematical models | en_ZA |
| dc.subject.lcsh | Blood cells | en_ZA |
| dc.subject.lcsh | Plasmodium falciparum | en_ZA |
| dc.subject.lcsh | Malaria -- Mathematical models | en_ZA |
| dc.subject.lcsh | Enzymes -- Mechanism of action | en_ZA |
| dc.subject.lcsh | Multilevel models (Statistics) | en_ZA |
| dc.subject.lcsh | Whole body level | en_ZA |
| dc.subject.name | UCTD | en_ZA |
| dc.title | Computational and analytical methods for constructing a multilevel model for human glucose metabolism | en_ZA |
| dc.type | Thesis | en_ZA |
Files
Original bundle
1 - 1 of 1
Loading...
- Name:
- green_computational_2022.pdf
- Size:
- 11.1 MB
- Format:
- Adobe Portable Document Format
- Description: