The crystalline lens is a unique cellular organ that performs metabolic processes while maintaining optical functionality. Mitochondria play a vital role in providing the cell with the energy necessary for these metabolic processes and have recently been shown to be more metabolically active than previously thought. To test the hypothesis that mitochondrial function directly influences the optical function of the lens, bovine lenses were treated with 50 μM, 200 μM, 600 μM and 1000 μM menadione, a mitochondrial specific toxin that renders the mitochondria inactive, and the Back Vertex Distance (BVD) variability was observed over 216 hours. Confocal micrographs of secondary fibre cells’ mitochondria were also analyzed for 50 μM, 200 μM, and 600 μM menadione treatment over 48 hours. Increase in BVD variability (± s.e.m.) was observed within 24 hours from 0.28 ± 0.021 to 1.83 ± 0.75 for the 600 μM treated lenses. Confocal micrograph analysis showed a trend toward a decrease in the average length of mitochondria from 7.9 ± 0.8 to 3.7 ± 0.9 over for 200 μM treated lenses and from 5.9 ± 1.0 to 3.6 ± 0.6 for the 600 μM treated lenses over 48 hours. These data show that indeed menadione has a detrimental effect on mitochondria as a function of both time and concentration and this change in mitochondria precedes changes in BVD variability directly linking mitochondrial function to optical function.

The crystalline lens is a unique cellular organ that performs metabolic processes while maintaining optical functionality. Mitochondria play a vital role in providing the cell with the energy necessary for these metabolic processes and have recently been shown to be more metabolically active than previously thought. To test the hypothesis that mitochondrial function directly influences the optical function of the lens, bovine lenses were treated with 50 μM, 200 μM, 600 μM and 1000 μM menadione, a mitochondrial specific toxin that renders the mitochondria inactive, and the Back Vertex Distance (BVD) variability was observed over 216 hours. Confocal micrographs of secondary fibre cells’ mitochondria were also analyzed for 50 μM, 200 μM, and 600 μM menadione treatment over 48 hours. Increase in BVD variability (± s.e.m.) was observed within 24 hours from 0.28 ± 0.021 to 1.83 ± 0.75 for the 600 μM treated lenses. Confocal micrograph analysis showed a trend toward a decrease in the average length of mitochondria from 7.9 ± 0.8 to 3.7 ± 0.9 over for 200 μM treated lenses and from 5.9 ± 1.0 to 3.6 ± 0.6 for the 600 μM treated lenses over 48 hours. These data show that indeed menadione has a detrimental effect on mitochondria as a function of both time and concentration and this change in mitochondria precedes changes in BVD variability directly linking mitochondrial function to optical function.

Optic nerve section (ONS) is an experimental model for damage of the optic nerve associated with diseases such as glaucoma and optic neuritis. Damage to the optic nerve causes loss of retinal ganglion cells that are attached, once the cells are damaged, they are not typically replaced. Recently, Fischer and Reh (2003) found that Müller glia have the potential to adopt phenotypes and functional capabilities similar to those of retinal progenitors, a potential source of retinal regeneration. In the chick, when the specific retinal cells are targeted for damage by chemotoxins, there is widespread apoptosis but also mitotically active cells that label with retinal progenitor markers. Fischer and Reh (2002) also discovered that the combination of growth factors FGF2 and insulin is capable of stimulating the regenerative response of the Müller glia to retinal progenitor cells in chick eyes. This study was conducted to analyse damage to the ganglion cells by optic nerve section in chicks to determine the effect of age on the cell death timeline, the proliferative qualities of the retina and to see if injections of growth factors had the ability to increase the proliferation. Histological methods were used to analyse cellular changes and ultrasound to monitor eye growth. Apoptotic activity preceded retinal thinning and ganglion cell loss, indicating that ONS-related cell death is mediated at least in part by apoptotic mechanisms and age did not affect the time course, although, age did affect the eye growth changes, which may be attributed to the plasticity of the younger eyes. ONS damage elicited proliferative activity in the retina as did growth factor injections alone. The combination of ONS damage and growth factor injections increased the proliferative activity and the overall total number of cells in the ganglion cell layer. These findings can potentially lead to the development of therapeutic strategies for the preservation or restoration of retinal cells in diseased eyes.

To paraphrase Stan Ulam, a Polish mathematician who became a leading figure in the Manhattan Project, in this dissertation I focus not only on how computer science can help biologists, but also on how biology can inspire computer scientists. On one hand, computer science provides powerful abstraction tools for metabolic networks. Cell metabolism is the set of chemical reactions taking place in a cell, with the aim of maintaining the living state of the cell. Due to the intrinsic complexity of metabolic networks, predicting the phenotypic traits resulting from a given genotype and metabolic structure is a challenging task. To this end, mathematical models of metabolic networks, called genome-scale metabolic models, contain all known metabolic reactions in an organism and can be analyzed with computational methods. In this dissertation, I propose a set of methods to investigate models of metabolic networks. These include multi-objective optimization, sensitivity, robustness and identifiability analysis, and are applied to a set of genome-scale models. Then, I augment the framework to predict metabolic adaptation to a changing environment. The adaptation of a microorganism to new environmental conditions involves shifts in its biochemical network and in the gene expression level. However, gene expression profiles do not provide a comprehensive understanding of the cellular behavior. Examples are the cases in which similar profiles may cause different phenotypic outcomes, while different profiles may give rise to similar behaviors. In fact, my idea is to study the metabolic response to diverse environmental conditions by predicting and analyzing changes in the internal molecular environment and in the underlying multi-omic networks. I also adapt statistical and mathematical methods (including principal component analysis and hypervolume) to evaluate short term metabolic evolution and perform comparative analysis of metabolic conditions. On the other hand, my vision is that a biomolecular system can be cast as a ?biological computer?, therefore providing insights into computational processes. I therefore study how computation can be performed in a biological system by proposing a map between a biological organism and the von Neumann architecture, where metabolism executes reactions mapped to instructions of a Turing machine. A Boolean string represents the genetic knockout strategy and also the executable program stored in the ?memory? of the organism. I use this framework to investigate scenarios of communication among cells, gene duplication, and lateral gene transfer. Remarkably, this mapping allows estimating the computational capability of an organism, taking into account also transmission events and communication outcomes.

In this thesis, we consider the problem of long-term persistence in ecological and epidemiological systems. This is important in conservation biology for protecting species at risk of extinction and in epidemiology for reducing disease prevalence and working towards elimination. Understanding how to predict and control persistence is critical for these aims. In Chapter 2, we discuss existing ways of characterising persistence and their relationship with the modelling paradigms employed in ecology and epidemiology. We note that data are often limited to information on the state of particular patches or populations and are modelled using a metapopulation approach. In Chapter 3, we define persistence in relation to a pre-specified time horizon in stochastic single-species and two-species competition models, comparing results between discrete and continuous time simulations. We find that discrete and continuous time simulations can result in different persistence predictions, especially in the case of inter-specific competition. The study also serves to illustrate the shortcomings of defining persistence in relation to a specific time horizon. A more mathematically rigorous interpretation of persistence in stochastic models can be found by considering the quasi-stationary distribution (QSD) and the associated measure of mean time to extinction from quasi-stationarity. In Chapter 4, we investigate the contribution of individual patches to extinction times and metapopulation size, and provide predictors of patch value that can be calculated easily from readily available data. In Chapter 5, we focus directly on the QSD of heterogeneous systems. Through simulation, we investigate possible compressions of the QSD that could be used when standard numerical approaches fail due to high system dimensionality, and provide guidance on appropriate compression choices for different purposes. In Chapter 6, we consider deterministic models and investigate the effect of introducing additional patch states on the persistence threshold. We suggest a possible model that might be appropriate for making predictions that extend to stochastic systems. By considering a family of models as limiting cases of a more general model, we demonstrate a novel approach for deriving quantities of interest for linked models that should help guide modelling decisions. Finally, in Chapter 7, we draw out implications for conservation biology and disease control, as well as for future work on biological persistence.

Motivation: A rapid technological development in the biosciences and in computer science in the last decade has enabled the analysis of high-dimensional biological datasets on standard desktop computers. However, in spite of these technical advances, common properties of the new high-throughput experimental data, like small sample sizes in relation to the number of features, high noise levels and outliers, also pose novel challenges. Ensemble and consensus machine learning techniques and data integration methods can alleviate these issues, but often provide overly complex models which lack generalization capability and interpretability. The goal of this thesis was therefore to develop new approaches to combine algorithms and large-scale biological datasets, including novel approaches to integrate analysis types from different domains (e.g. statistics, topological network analysis, machine learning and text mining), to exploit their synergies in a manner that provides compact and interpretable models for inferring new biological knowledge. Main results: The main contributions of the doctoral project are new ensemble, consensus and cross-domain bioinformatics algorithms, and new analysis pipelines combining these techniques within a general framework. This framework is designed to enable the integrative analysis of both large- scale gene and protein expression data (including the tools ArrayMining, Top-scoring pathway pairs and RNAnalyze) and general gene and protein sets (including the tools TopoGSA , EnrichNet and PathExpand), by combining algorithms for different statistical learning tasks (feature selection, classification and clustering) in a modular fashion. Ensemble and consensus analysis techniques employed within the modules are redesigned such that the compactness and interpretability of the resulting models is optimized in addition to the predictive accuracy and robustness. The framework was applied to real-word biomedical problems, with a focus on cancer biology, providing the following main results: (1) The identification of a novel tumour marker gene in collaboration with the Nottingham Queens Medical Centre, facilitating the distinction between two clinically important breast cancer subtypes (framework tool: ArrayMining) (2) The prediction of novel candidate disease genes for Alzheimer’s disease and pancreatic cancer using an integrative analysis of cellular pathway definitions and protein interaction data (framework tool: PathExpand, collaboration with the Spanish National Cancer Centre) (3) The prioritization of associations between disease-related processes and other cellular pathways using a new rule-based classification method integrating gene expression data and pathway definitions (framework tool: Top-scoring pathway pairs) (4) The discovery of topological similarities between differentially expressed genes in cancers and cellular pathway definitions mapped to a molecular interaction network (framework tool: TopoGSA, collaboration with the Spanish National Cancer Centre) In summary, the framework combines the synergies of multiple cross-domain analysis techniques within a single easy-to-use software and has provided new biological insights in a wide variety of practical settings.

Since the completion of the Human Genome Project in 2003, it has become increasingly apparent that while genomics has a major role to play in the understanding of human biology, information from other disciplines is necessary to explain the web of interacting signals that allow our bodies to function on a day to day basis and respond to rapid changes in our local environment. One such field, that of metabolomics, focuses on the study of the set of low molecular weight compounds (metabolites) involved in metabolism. Metabolomic studies aim to quantify the concentrations of each of these compounds within a subject under particular conditions, resulting in either information on the physiological effects of a disease or environmental factor (such as a toxin) on the organism, or the identification of metabolites or groups of metabolites that serve as biochemical markers for a state or illness. Whilst metabolite concentrations alone can give great insight into a chosen state, additional information can be obtained by considering the ways in which metabolites interact with each other as parts of a larger system. One method of tackling this problem, metabolic networks, is gaining popularity within the community as it offers a complementary approach to the traditional biological method for studying metabolism, the metabolic pathway. Construction methods are varied; ranging from the mapping of experimental data onto pathway diagrams, through the use of correlation-based techniques, to the analysis of time-series data of metabolic fluxes. However, while many attempts have been made to capture and visualise the complex web of reactions within an organism, few have yet succeeded in showing how they can be used to help identify the metabolites that are most significantly involved in the differences between groups of biological samples. This thesis discusses ways in which graphs may be used to aid researchers in both the visualisation and interpretation of metabolomic datasets, and provide a platform for more automated analysis techniques. To that end, it first presents the background to the relevant topics, metabolomics and graph theory, before moving on to show how metabolic correlation networks can be used to identify and visualise differences in metabolism between groups of subjects. It then introduces Linked Metabolites, a software package that has been developed to help researchers explain differences in metabolism by highlighting relationships between metabolites within the metabolic pathways, and to compile those relationships into directed metabolic graphs suitable for analysis using metrics from graph theory. Finally, the thesis explains how the directed metabolic graphs produced by Linked Metabolites could potentially be used to integrate data gathered from the same sample using different experimental techniques, refining the areas of the underlying biochemistry needing further investigation.

The aim of this work is to provide geneticists with appropriate statistical methods and computer programmes for the analysis of human pedigree data in view of mapping genes on the human chromosomes, and discovering the origin of chromosomal abnormalities such as the autosomal trisomies, the 47,XXY Klinefelter's syndrome and the 46,XX men syndrome. J.H. Edwards' marker algebra is presented in detail as used in his computer programme (MARK III) that analyses linkage with Morton's lod method for normal diploids. The programme is also described with all its specifications. The cytological mechanisms leading to autosomal trisomy are described to show that the proportion of trisomics carrying three alleles from three of their grandparents is bound to be greater than zero for any locus anywhere on a trisomic chromosome. The use of A.W.F. Edwards' method of support is then demonstrated on various sets of data to definitely exclude the ABO, MN, P, Jk, Gc and Lp loci from chromosome no. 21, and the theory is extended to show that about 401 of 47,XXY men receive an extra X from their fathers and 60% from their mothers, and that in general 46,XX men are more likely to arise from 47,XXY zygotes that lose their Y chromosomes than by an interchange between the X and Y chromosomes of their fathers.

This thesis addresses the problem of active module identification in biological networks. Active module identification is a research topic in network biology that aims to identify regions in network showing striking changes in activity. It is often associated with a given cellular response and expected to reveal dynamic and process-specific information. The key research questions for this thesis are the practical formulations of active module identification problem,the design of effective, efficient and robust algorithms to identify active modules, and the right way to interpret identified active module. This thesis contributes by proposing three different algorithm frameworks to address the research question from three different aspects. It first explores an integrated approach of combining both gene differential expression and differential correlation, formulates it as a multi-objective problem, and solves it on both simulated data and real world data. Then the thesis investigates a novel approach that brings in prior knowledge of biological process, and balances between pure data-driven search and prior information guidance. Finally, the thesis presents a brand new framework of identifying active module and topological communities simultaneously using evolutionary multitasking, accompanied with a series of task-specific algorithm designs and improvements, and provides a new way of integrating topological information to help the interpretation of active module.

Myopia is a common ocular condition worldwide and the mechanism of myopia is still not clear. A number of animal models of myopia and refractive error development have been proposed. The fact that form deprivation myopia could be induced in tilapia fish, as shown previously in my research, suggests the possibility that tilapia could be a new animal model for myopia research. In the first part of this thesis the tilapia model was perfected and then, based on this model, the effect of systemic hormones (thyroid hormones) associated with eye and body development was investigated during refractive error development. Lastly, the physiological and morphological changes on the retina were further studied with optical coherence tomography (OCT). In these experiments, significant amounts of myopia, and hyperopia were induced within two weeks using goggles with lens inserts as in other higher vertebrate animal models, e.g. chicks. The results from form deprivation treatment also show that the sensitivity of tilapia eyes may be an age related effect during the emmetropization process. The larger the fish, the less hyperopic the fish eye, though the small eye artefact may be a factor. The susceptibility of the refractive development of the eye to the visual environment may be also linked to plasma hormone levels. It was found that induced refractive errors could be shifted in the hyperopic direction with high levels of thyroid hormones. Also, after 2 weeks of treatment with negative or positive lens/goggles, the tilapia retina becomes thinner or thicker, respectively. When the goggles are removed, the thickness of the retina changes within hours and gradually returns to normal. However, the circadian retinomotor movement is a complicating factor since it affects the retinal thickness measurement with OCT at some time points. In conclusion, tilapia represent a good lower vertebrate model for myopia research, suggesting a universal mechanism of myopia development, which may involve systemic hormones and immediate, short term retinal responses.