Use this URL to cite or link to this record in EThOS:
Title: A multiscale discrete model integration strategy for Systems Biology implemented in a grid-enabled software platform : an example application from cancer systems modelling
Author: Patel, Manish
ISNI:       0000 0004 2668 2564
Awarding Body: UCL (University College London)
Current Institution: University College London (University of London)
Date of Award: 2008
Availability of Full Text:
Access from EThOS:
Full text unavailable from EThOS. Please try the link below.
Access from Institution:
Model integration - the process by which different modelling efforts can be brought together to simulate the target system - is a core technology in the field of Systems Biology. In the work presented here model integration was addressed directly taking cancer systems as an example. An in-depth literature review was carried out to survey the model forms and types currently being utilised. This was used to formalise the main challenges that model integration poses, namely that of paradigm (the formalism on which a model is based), focus (the real-world system the model represents) and scale. A two-tier model integration strategy, including a knowledge-driven approach to address model semantics, was developed to tackle these challenges. In the first step a novel description of models at the level of behaviour, rather than the precise mathematical or computational basis of the model, is developed by distilling a set of abstract classes and properties. These can accurately describe model behaviour and hence describe focus in a way that can be integrated with behavioural descriptions of other models. In the second step this behaviour is decomposed into an agent-based system by translating the models into local interaction rules. These rules must be enriched and the agent model simulated, therefore a Grid-like Java infrastructure was developed and tested on an 18-node Beowulf cluster. The two-tier approach was tested on this software by taking 4 different models, each exhibiting complexities and submodels, that were successfully integrated and simulated together. The results showed all of the main challenges could be overcome given the correct conditions for rule enrichment, in this case implemented as a genetic algorithm that operated on rule components. This research represents a key breakthrough for cancer systems research. The two-tier approach could provide the tools necessary to understand tumour behavioural complexity and hence provide a means to combat the disease.
Supervisor: Not available Sponsor: Not available
Qualification Name: Thesis (Ph.D.) Qualification Level: Doctoral
EThOS ID:  DOI: Not available