MicroRNAs (miRNAs) are potent effectors in gene regulatory networks where aberrant miRNA expression can contribute to human diseases such as cancer. For a better understanding of the regulatory role of miRNAs in coordinating gene expression, we here present a systems biology approach combining data-driven modeling and model-driven experiments. Such an approach is characterized by an iterative process, including biological data acquisition and integration, network construction, mathematical modeling and experimental validation. To demonstrate the application of this approach, we adopt it to investigate mechanisms of collective repression on p21 by multiple miRNAs. We first construct a p21 regulatory network based on data from the literature and further expand it using algorithms that predict molecular interactions. Based on the network structure, a detailed mechanistic model is established and its parameter values are determined using data. Finally, the calibrated model is used to study the effect of different miRNA expression profiles and cooperative target regulation on p21 expression levels in different biological contexts.
Phosphoglycerate-mutase (PGM) is an ubiquitous glycolytic enzyme, which in eukaryotic cells can be found in different compartments. In prokaryotic cells, several PGMs are annotated/localized in one compartment. The identification and functional characterization of PGMs in prokaryotes is therefore important for better understanding of metabolic regulation. Here we introduce a method, based on a multi-level kinetic model of the primary carbon metabolism in cyanobacterium Synechococcus elongatus PCC 7942, that allows the identification of a specific function for a particular PGM. The strategy employs multiple parameter estimation runs in high CO2, combined with simulations testing a broad range of kinetic parameters against the changes in transcript levels of annotated PGMs. Simulations are evaluated for a match in metabolic level in low CO2, to reveal trends that can be linked to the function of a particular PGM. A one-isoenzyme scenario shows that PGM2 is a major regulator of glycolysis, while PGM1 and PGM4 make the system robust against environmental changes. Strikingly, combining two PGMs with reverse transcriptional regulation allows both features. A conclusion arising from our analysis is that a two-enzyme PGM system is required to regulate the flux between glycolysis and the Calvin-Benson cycle, while an additional PGM increases the robustness of the system.
The stressosome is a bacterial signalling complex that responds to environmental changes by initiating a protein partner switching cascade, which leads to the release of the alternative sigma factor, σB. Stress perception increases the phosphorylation of the stressosome sensor protein, RsbR, and the scaffold protein, RsbS, by the protein kinase, RsbT. Subsequent dissociation of RsbT from the stressosome activates the σB cascade. However, the sequence of physical events that occur in the stressosome during signal transduction is insufficiently understood.
Here, we use computational modelling to correlate the structure of the stressosome with the efficiency of the phosphorylation reactions that occur upon activation by stress. In our model, the phosphorylation of any stressosome protein is dependent upon its nearest neighbours and their phosphorylation status. We compare different hypotheses about stressosome activation and find that only the model representing the allosteric activation of the kinase RsbT, by phosphorylated RsbR, qualitatively reproduces the experimental data.
Our simulations and the associated analysis of published data support the following hypotheses: (i) a simple Boolean model is capable of reproducing stressosome dynamics, (ii) different stressors induce identical stressosome activation patterns, and we also confirm that (i) phosphorylated RsbR activates RsbT, and (ii) the main purpose of RsbX is to dephosphorylate RsbS-P.
Bacillus subtilis; Stressosome; Signalling; Cellular automaton; Stress response
The present work exemplifies how parameter identifiability analysis can be used to gain insights into differences in experimental systems and how uncertainty in parameter estimates can be handled. The case study, presented here, investigates interferon-gamma (IFNγ) induced STAT1 signalling in two cell types that play a key role in pancreatic cancer development: pancreatic stellate and cancer cells. IFNγ inhibits the growth for both types of cells and may be prototypic of agents that simultaneously hit cancer and stroma cells. We combined time-course experiments with mathematical modelling to focus on the common situation in which variations between profiles of experimental time series, from different cell types, are observed. To understand how biochemical reactions are causing the observed variations, we performed a parameter identifiability analysis. We successfully identified reactions that differ in pancreatic stellate cells and cancer cells, by comparing confidence intervals of parameter value estimates and the variability of model trajectories. Our analysis shows that useful information can also be obtained from nonidentifiable parameters. For the prediction of potential therapeutic targets we studied the consequences of uncertainty in the values of identifiable and nonidentifiable parameters. Interestingly, the sensitivity of model variables is robust against parameter variations and against differences between IFNγ induced STAT1 signalling in pancreatic stellate and cancer cells. This provides the basis for a prediction of therapeutic targets that are valid for both cell types.
For the prediction of therapeutic targets and the design of therapies, it is important to study the same pathway across different cell types. This is particularly relevant for cancer research, where several cell types are involved in carcinogenesis. Pancreatic cancer is enhanced by activated pancreatic stellate cells. It would thus seem plausible for an effective therapy to hit stellate and cancer cells. The cytokine IFNγ is an inhibitor of proliferation in both cell types. Antiproliferative effects of IFNγ are mediated by STAT1 signalling. An important aspect is to determine those reactions that cause the differences in the initial increase of phosphorylated STAT1 and in the temporal profile of STAT1 nuclear accumulation between the two cell types. We examined this aspect by performing a parameter identifiability analysis for calibrated mathematical models. We calculated confidence intervals of the estimated parameter values and found that they provide insights into reactions underlying the differences. A key finding of sensitivity analysis elucidated that predicted targets for enhancement of STAT1 activity are robust against parameter uncertainty and moreover they are robust between the two cell types. Our case study therefore exemplified how identifiability and sensitivity analysis can provide a basis for the prediction of potential therapeutic targets.
AIM: To gain insights into the molecular action of erlotinib in pancreatic cancer (PC) cells.
METHODS: Two PC cell lines, BxPC-3 and Capan-1, were treated with various concentrations of erlotinib, the specific mitogen-activated protein kinase kinase (MEK) inhibitor U0126, and protein kinase B (AKT) inhibitor XIV. DNA synthesis was measured by 5-bromo-2'-deoxyuridine (BrdU) assays. Expression and phosphorylation of the epidermal growth factor receptor (EGFR) and downstream signaling molecules were quantified by Western blot analysis. The data were processed to calibrate a mathematical model, based on ordinary differential equations, describing the EGFR-mediated signal transduction.
RESULTS: Erlotinib significantly inhibited BrdU incorporation in BxPC-3 cells at a concentration of 1 μmol/L, whereas Capan-1 cells were much more resistant. In both cell lines, MEK inhibitor U0126 and erlotinib attenuated DNA synthesis in a cumulative manner, whereas the AKT pathway-specific inhibitor did not enhance the effects of erlotinib. While basal phosphorylation of EGFR and extracellular signal-regulated kinase (ERK) did not differ much between the two cell lines, BxPC-3 cells displayed a more than five-times higher basal phospho-AKT level than Capan-1 cells. Epidermal growth factor (EGF) at 10 ng/mL induced the phosphorylation of EGFR, AKT and ERK in both cell lines with similar kinetics. In BxPC-3 cells, higher levels of phospho-AKT and phospho-ERK (normalized to the total protein levels) were observed. Independent of the cell line, erlotinib efficiently inhibited phosphorylation of EGFR, AKT and ERK. The mathematical model successfully simulated the experimental findings and provided predictions regarding phosphoprotein levels that could be verified experimentally.
CONCLUSION: Our data suggest basal AKT phosphorylation and the degree of EGF-induced activation of AKT and ERK as molecular determinants of erlotinib efficiency in PC cells.
Erlotinib; Pancreatic cancer; Epidermal growth factor receptor; Signal transduction; Mathematical modeling
MicroRNA (miRNA) target hubs are genes that can be simultaneously targeted by a comparatively large number of miRNAs, a class of non-coding RNAs that mediate post-transcriptional gene repression. Although the details of target hub regulation remain poorly understood, recent experiments suggest that pairs of miRNAs can cooperate if their binding sites reside in close proximity. To test this and other hypotheses, we established a novel approach to investigate mechanisms of collective miRNA repression. The approach presented here combines miRNA target prediction and transcription factor prediction with data from the literature and databases to generate a regulatory map for a chosen target hub. We then show how a kinetic model can be derived from the regulatory map. To validate our approach, we present a case study for p21, one of the first experimentally proved miRNA target hubs. Our analysis indicates that distinctive expression patterns for miRNAs, some of which interact cooperatively, fine-tune the features of transient and long-term regulation of target genes. With respect to p21, our model successfully predicts its protein levels for nine different cellular functions. In addition, we find that high abundance of miRNAs, in combination with cooperativity, can enhance noise buffering for the transcription of target hubs.
Proteolytic breakdown of the amyloid precursor protein (APP) by secretases is a complex cellular process that results in formation of neurotoxic Aβ peptides, causative of neurodegeneration in Alzheimer’s disease (AD). Processing involves monomeric and dimeric forms of APP that traffic through distinct cellular compartments where the various secretases reside. Amyloidogenic processing is also influenced by modifiers such as sorting receptor-related protein (SORLA), an inhibitor of APP breakdown and major AD risk factor.
In this study, we developed a multi-compartment model to simulate the complexity of APP processing in neurons and to accurately describe the effects of SORLA on these processes. Based on dose–response data, our study concludes that SORLA specifically impairs processing of APP dimers, the preferred secretase substrate. In addition, SORLA alters the dynamic behavior of β-secretase, the enzyme responsible for the initial step in the amyloidogenic processing cascade.
Our multi-compartment model represents a major conceptual advance over single-compartment models previously used to simulate APP processing; and it identified APP dimers and β-secretase as the two distinct targets of the inhibitory action of SORLA in Alzheimer’s disease.
Amyloidogenic processing; Compartmental modeling; LR11; Secretases; SORL1; VPS10P domain receptors
Ageing is a complex multifactorial process involving a progressive physiological decline that, ultimately, leads to the death of an organism. It involves multiple changes in many components that play fundamental roles under healthy and pathological conditions. Simultaneously, every organism undergoes accumulative ‘wear and tear’ during its lifespan, which confounds the effects of the ageing process. The scenario is complicated even further by the presence of both age-dependent and age-independent competing causes of death. Various manipulations have been shown to interfere with the ageing process. Calorie restriction, for example, has been reported to increase the lifespan of a wide range of organisms, which suggests a strong relation between energy metabolism and ageing. Such a link is also supported within the main theories for ageing: the free radical hypothesis, for instance, links oxidative damage production directly to energy metabolism. The Dynamic Energy Budgets (DEB) theory, which characterizes the uptake and use of energy by living organisms, therefore constitutes a useful tool for gaining insight into the ageing process. Here we compare the existing DEB-based modelling approaches and, then, discuss how new biological evidence could be incorporated within a DEB framework.
metabolism; senescence; whole-body energetics; calorie restriction, systems biology
Modeling the Calvin-Benson cycle has a history in the field of theoretical biology. Anyone who intends to model this system will look at existing models to adapt, refine and improve them. With the goal to study the regulation of carbon metabolism, we investigated a broad range of relevant models for their suitability to provide the basis for further modeling efforts. Beyond a critical analysis of existing models, we furthermore investigated the question how adjacent metabolic pathways, for instance photorespiration, can be integrated in such models.
Our analysis reveals serious problems with a range of models that are publicly available and widely used. The problems include the irreproducibility of the published results or significant differences between the equations in the published description of the model and model itself in the supplementary material. In addition to and based on the discussion of existing models, we furthermore analyzed approaches in PGA sink implementation and confirmed a weak relationship between the level of its regulation and efficiency of PGA export, in contrast to significant changes in the content of metabolic pool within the Calvin-Benson cycle.
In our study we show that the existing models that have been investigated are not suitable for reuse without substantial modifications. We furthermore show that the minor adjacent pathways of the carbon metabolism, neglected in all kinetic models of Calvin-Benson cycle, cannot be substituted without consequences in the mass production dynamics. We further show that photorespiration or at least its first step (O2 fixation) has to be implemented in the model if this model is aimed for analyses out of the steady state.
When analyzing microarray data, non-biological variation introduces uncertainty in the analysis and interpretation. In this paper we focus on the validation of significant differences in gene expression levels, or normalized channel intensity levels with respect to different experimental conditions and with replicated measurements. A myriad of methods have been proposed to study differences in gene expression levels and to assign significance values as a measure of confidence. In this paper we compare several methods, including SAM, regularized t-test, mixture modeling, Wilk’s lambda score and variance stabilization. From this comparison we developed a weighted resampling approach and applied it to gene deletions in Mycobacterium bovis.
We discuss the assumptions, model structure, computational complexity and applicability to microarray data. The results of our study justified the theoretical basis of the weighted resampling approach, which clearly outperforms the others.
Algorithms were implemented using the statistical programming language R and available on the author’s web-page.
Interferon-gamma (IFNγ) is a multifunctional cytokine with antifibrotic and antiproliferative efficiency. We previously found that pancreatic stellate cells (PSC), the main effector cells in cancer-associated fibrosis, are targets of IFNγ action in the pancreas. Applying a combined experimental and computational approach, we have demonstrated a pivotal role of STAT1 in IFNγ signaling in PSC. Using in vivo and in vitro models of pancreatic cancer, we have now studied IFNγ effects on the tumor cells themselves. We hypothesize that IFNγ inhibits tumor progression through two mechanisms, reduction of fibrogenesis and antiproliferative effects on the tumor cells. To elucidate the molecular action of IFNγ, we have established a mathematical model of STAT1 activation and combined experimental studies with computer simulations.
In BALB/c-nu/nu mice, flank tumors composed of DSL-6A/C1 pancreatic cancer cells and PSC grew faster than pure DSL-6A/C1 cell tumors. IFNγ inhibited the growth of both types of tumors to a similar degree. Since the stroma reaction typically reduces the efficiency of therapeutic agents, these data suggested that IFNγ may retain its antitumor efficiency in PSC-containing tumors by targeting the stellate cells. Studies with cocultures of DSL-6A/C1 cells and PSC revealed a modest antiproliferative effect of IFNγ under serum-free conditions. Immunoblot analysis of STAT1 phosphorylation and confocal microscopy studies on the nuclear translocation of STAT1 in DSL-6A/C1 cells suggested that IFNγ-induced activation of the transcription factor was weaker than in PSC. The mathematical model not only reproduced the experimental data, but also underscored the conclusions drawn from the experiments by indicating that a maximum of 1/500 of total STAT1 is located as phosphorylated STAT1 in the nucleus upon IFNγ treatment of the tumor cells.
IFNγ is equally effective in DSL-6A/C1 tumors with and without stellate cells. While its action in the presence of PSC may be explained by inhibition of fibrogenesis, its efficiency in PSC-free tumors is unlikely to be caused by direct effects on the tumor cells alone but may involve inhibitory effects on local stroma cells as well. To gain further insights, we also plan to apply computer simulations to the analysis of tumor growth in vivo.
Clostridium acetobutylicum is an anaerobic bacterium which is known for its solvent-producing capabilities, namely regarding the bulk chemicals acetone and butanol, the latter being a highly efficient biofuel. For butanol production by C. acetobutylicum to be optimized and exploited on an industrial scale, the effect of pH-induced gene regulation on solvent production by C. acetobutylicum in continuous culture must be understood as fully as possible.
We present an ordinary differential equation model combining the metabolic network governing solvent production with regulation at the genetic level of the enzymes required for this process. Parameterizing the model with experimental data from continuous culture, we demonstrate the influence of pH upon fermentation products: at high pH (pH 5.7) acids are the dominant product while at low pH (pH 4.5) this switches to solvents. Through steady-state analyses of the model we focus our investigations on how alteration in gene expression of C. acetobutylicum could be exploited to increase butanol yield in a continuous culture fermentation.
Incorporating gene regulation into the model of solvent production by C. acetobutylicum enables an accurate representation of the pH-induced switch to solvent production to be obtained and theoretical investigations of possible synthetic-biology approaches to be pursued. Steady-state analyses suggest that, to increase butanol yield, alterations in the expression of single solvent-associated genes are insufficient; a more complex approach targeting two or more genes is required.
Multilevelness is a defining characteristic of complex systems. For example, in the intestinal tissue the epithelial lining is organized into crypts that are maintained by a niche of stem cells. The behavior of the system 'as a whole' is considered to emerge from the functioning and interactions of its parts. What we are seeking here is a conceptual framework to demonstrate how the "fate" of intestinal crypts is an emergent property that inherently arises from the complex yet robust underlying biology of stem cells.
We establish a conceptual framework in which to formalize cross-level principles in the context of tissue organization. To this end we provide a definition for stemness, which is the propensity of a cell lineage to contribute to a tissue fate. We do not consider stemness a property of a cell but link it to the process in which a cell lineage contributes towards tissue (mal)function. We furthermore show that the only logically feasible relationship between the stemness of cell lineages and the emergent fate of their tissue, which satisfies the given criteria, is one of dominance from a particular lineage.
The dominance theorem, conceived and proven in this paper, provides support for the concepts of niche succession and monoclonal conversion in intestinal crypts as bottom-up relations, while crypt fission is postulated to be a top-down principle.
Targeted therapy approaches have been successfully introduced into the treatment of several cancers. The multikinase inhibitor Sorafenib has antitumor activity in solid tumors and its effects on acute lymphoblastic leukemia (ALL) cells are still unclear.
ALL cell lines (SEM, RS4;11 and Jurkat) were treated with Sorafenib alone or in combination with cytarabine, doxorubicin or RAD001. Cell count, apoptosis and necrosis rates, cell cycle distribution, protein phosphorylation and metabolic activity were determined.
Sorafenib inhibited the proliferation of ALL cells by cell cycle arrest accompanied by down-regulation of CyclinD3 and CDK4. Furthermore, Sorafenib initiated apoptosis by cleavage of caspases 3, 7 and PARP. Apoptosis and necrosis rates increased significantly with most pronounced effects after 96 h. Antiproliferative effects of Sorafenib were associated with a decreased phosphorylation of Akt (Ser473 and Thr308), FoxO3A (Thr32) and 4EBP-1 (Ser65 and Thr70) as early as 0.5 h after treatment. Synergistic effects were seen when Sorafenib was combined with other cytotoxic drugs or a mTOR inhibitor emphasizing the Sorafenib effect.
Sorafenib displays significant antileukemic activity in vitro by inducing cell cycle arrest and apoptosis. Furthermore, it influences PI3K/Akt/mTOR signaling in ALL cells.
Non-coding RNAs gain more attention as their diverse roles in many cellular processes are discovered. At the same time, the need for efficient computational prediction of ncRNAs increases with the pace of sequencing technology. Existing tools are based on various approaches and techniques, but none of them provides a reliable ncRNA detector yet. Consequently, a natural approach is to combine existing tools. Due to a lack of standard input and output formats combination and comparison of existing tools is difficult. Also, for genomic scans they often need to be incorporated in detection workflows using custom scripts, which decreases transparency and reproducibility.
We developed a Java-based framework to integrate existing tools and methods for ncRNA detection. This framework enables users to construct transparent detection workflows and to combine and compare different methods efficiently. We demonstrate the effectiveness of combining detection methods in case studies with the small genomes of Escherichia coli, Listeria monocytogenes and Streptococcus pyogenes. With the combined method, we gained 10% to 20% precision for sensitivities from 30% to 80%. Further, we investigated Streptococcus pyogenes for novel ncRNAs. Using multiple methods--integrated by our framework--we determined four highly probable candidates. We verified all four candidates experimentally using RT-PCR.
We have created an extensible framework for practical, transparent and reproducible combination and comparison of ncRNA detection methods. We have proven the effectiveness of this approach in tests and by guiding experiments to find new ncRNAs. The software is freely available under the GNU General Public License (GPL), version 3 at http://www.sbi.uni-rostock.de/moses along with source code, screen shots, examples and tutorial material.
The main conclusion is that systems biology approaches can indeed advance cancer research, having already proved successful in a very wide variety of cancer-related areas, and are likely to prove superior to many current research strategies. Major points include:
Systems biology and computational approaches can make important contributions to research and development in key clinical aspects of cancer and of cancer treatment, and should be developed for understanding and application to diagnosis, biomarkers, cancer progression, drug development and treatment strategies.Development of new measurement technologies is central to successful systems approaches, and should be strongly encouraged. The systems view of disease combined with these new technologies and novel computational tools will over the next 5–20 years lead to medicine that is predictive, personalized, preventive and participatory (P4 medicine).Major initiatives are in progress to gather extremely wide ranges of data for both somatic and germ-line genetic variations, as well as gene, transcript, protein and metabolite expression profiles that are cancer-relevant. Electronic databases and repositories play a central role to store and analyze these data. These resources need to be developed and sustained.Understanding cellular pathways is crucial in cancer research, and these pathways need to be considered in the context of the progression of cancer at various stages. At all stages of cancer progression, major areas require modelling via systems and developmental biology methods including immune system reactions, angiogenesis and tumour progression.A number of mathematical models of an analytical or computational nature have been developed that can give detailed insights into the dynamics of cancer-relevant systems. These models should be further integrated across multiple levels of biological organization in conjunction with analysis of laboratory and clinical data.Biomarkers represent major tools in determining the presence of cancer, its progression and the responses to treatments. There is a need for sets of high-quality annotated clinical samples, enabling comparisons across different diseases and the quantitative simulation of major pathways leading to biomarker development and analysis of drug effects.Education is recognized as a key component in the success of any systems biology programme, especially for applications to cancer research. It is recognized that a balance needs to be found between the need to be interdisciplinary and the necessity of having extensive specialist knowledge in particular areas.A proposal from this workshop is to explore one or more types of cancer over the full scale of their progression, for example glioblastoma or colon cancer. Such an exemplar project would require all the experimental and computational tools available for the generation and analysis of quantitative data over the entire hierarchy of biological information. These tools and approaches could be mobilized to understand, detect and treat cancerous processes and establish methods applicable across a wide range of cancers.
Systems biology; EU-USA workshop; Cancer
In recent years it has become evident that ABC transporters fulfill important barrier functions in normal organs and during disease processes. Most importantly, resistance to drugs in cancer cells led to intense oncological and pharmacological investigations in which researchers were able to highlight important pharmacological interactions of chemotherapeuticals with ABC transporter function. Recently, the development of neurodegenerative diseases and the maintenance of neuronal stem cells have been linked to the activity of ABC transporters. Here, we summarize findings from cell culture experiments, animal models and studies of patients with Alzheimer’s disease. Furthermore, we discuss pharmacological interactions and computational methods for risk assessment.
ABC transporter; Alzheimer’s disease; neurodegeneration; MDR1; MRP1; BCRP; systems biology; dementia; Abeta; blood-brain barrier; p-glycoprotein
Signalling pathways are complex systems in which not only simple monomeric molecules interact, but also more complex structures that include constitutive or induced protein assemblies. In particular, the hetero-and homo-dimerisation of proteins is a commonly encountered motif in signalling pathways. Several authors have suggested in recent times that dimerisation relates to a series of physical and biological outcomes used by the cell in the regulation of signal transduction.
In this paper we investigate the role of homodimerisation in receptor-protein transducer interactions. Towards this end, mathematical modelling is used to analyse the features of such kind of interactions and to predict the behaviour of the system under different experimental conditions. A kinetic model in which the interaction between homodimers provokes a dual mechanism of activation (single and double protein transducer activation at the same time) is proposed. In addition, we analyse under which conditions the use of a power-law representation for the system is useful. Furthermore, we investigate the dynamical consequences of this dual mechanism and compare the performance of the system in different simulated experimental conditions.
The analysis of our mathematical model suggests that in receptor-protein interacting systems with dual mechanism there may be a shift between double and single activation in a way that intense double protein transducer activation could initiate and dominate the signal in the short term (getting a fast intense signal), while single protein activation could control the system in the medium and long term (when input signal is weaker and decreases slowly). Our investigation suggests that homodimerisation and oligomerisation are mechanisms used to enhance and regulate the dynamic properties of the initial steps in signalling pathways.
The amplification of signals, defined as an increase in the intensity of a signal through networks of intracellular reactions, is considered one of the essential properties in many cell signalling pathways. Despite of the apparent importance of signal amplification, there have been few attempts to formalise this concept.
In this work we investigate the amplification and responsiveness of the JAK2-STAT5 pathway using a kinetic model. The recruitment of EpoR to the plasma membrane, activation by Epo, and deactivation of the EpoR/JAK2 complex are considered as well as the activation and nucleocytoplasmic shuttling of STAT5. Using qualitative biological knowledge, we first establish the structure of a general power-law model. We then generate a family of models from which we select suitable candidates. The parameter values of the model are estimated from experimental quantitative time-course data. The final model, whether it is conventional model with fixed predefined integer kinetic orders or a model with variable non-integer kinetic orders, is selected on the basis of a good agreement between simulations and the experimental data. The model is used to analyse the responsiveness and amplification properties of the pathway with sustained, transient, and oscillatory stimulation.
The selected kinetic model predicts that the system acts as an amplifier with maximum amplification and sensitivity for input signals whose intensity match physiological values for Epo concentration and with duration in the range of one to 100 minutes. The response of the system reaches saturation for more intense and longer stimulation with Epo. We hypothesise that these properties of the system directly relate to the saturation of Epo receptor activation, its low recruitment to the plasma membrane and intense deactivation as predicted by the model.
Despite its enormous promise to further our understanding of cellular processes involved in the regulation of gene expression, microarray technology generates data
for which statistical pre-processing has become a necessity before any interpretation
of data can begin. The process by which we distinguish (and remove) non-biological
variation from biological variation is called normalization. With a multitude of
experimental designs, techniques and technologies influencing the acquisition of data,
numerous approaches to normalization have been proposed in the literature. The
purpose of this short review is not to add to the many suggestions that have been
made, but to discuss some of the difficulties we encounter when analysing microarray