Tuesday, February 19, 2008

ICMSB 2008: Compendium of Abstracts

Modeling Biochemical Networks
Speaker: Eberhard Voit (UC Davis, USA)


Efficient Inferring Method of Genetic Interactions Based on Time Series Gene Expression Profiles: Application of Conceptual Modeling by S-system Formalism
Speaker: Masahiro Okamoto (Kyushu U, Japan)


The expression profiles of hundreds and thousands of genes on a genomic scale can be measured simultaneously by recent powerful technologies such as DNA microarrays, DNA chips and so on. These observed data depending on its environment are usually obtained as snapshots, but can be generated as dense time series that indicate the dynamic behavior. The experimentally observed time-course data should contain enormous information about the regulation of genetic networks in vivo. However, since this information is entirely implicit, it requires adequate analytical and computational methods of retrieval and interpretation. This inference problem of genetic networks based on the experimentally observed time-course data is generally referred to as "inverse problem" and can be defined as function optimization of the values of parameters involved in a suitable model-representation of genetic network. In brief, we have to estimate the values of a set of system parameters in the model which can realize given experimentally observed time-course data.

The key points solving such an inverse problem are how to set up canonical representation of mathematical modeling of genetic network and how to explore and exploit the values of parameters within immense huge searching space. We had first proposed a novel inferring method of genetic network by combining a dynamic network model called S-system with a computational technique of parameter estimation based on simple genetic algorithms. S-system is based on a particular type of differential equation in which the temporal (time-dependent) dynamic processes of system components are characterized by power-law formalism and S-system is suitable for the conceptual modeling and the description of organizationally complex systems involved looping or cyclic interactions between system components such as metabolic pathways and gene expression networks. The values of interrelated coefficient in this formalism are directly or indirectly related to the regulation mechanism in the modeled network, and the inferred network structure resulting from the estimation of parameters should be one of the better candidates for genetic network structure.

S-system formalism, however, has a major disadvantage in that this formalism includes a large number of parameters that must be estimated; the number of estimating parameters is 2n(n+1) (where n is the number of system components). Simple genetic algorithm (SGA) is one of the well-known heuristic optimizer of such large number of parameters, however, SGA has two intrinsic problems; one is early convergence in the first stage of search and the other is evolutionary stagnation in the last stage of search.

Recently real-coded genetic algorithms (RCGAs) attract attention as numerical optimizing methods instead of SGA. One of the crossover operators for RCGAs called unimodal normal distribution crossover (UNDX) has shown good performance in optimizing of various functions including multi-modal ones and benchmark functions with epistasis among parameters. Furthermore Sato et al. have proposed new generation-alternation of model called minimal generation gap (MGG) model to avoid early convergence in the first stage and to suppress evolutionary stagnation in the last

Using S-system modeling and RCGAs with the combination of the UNDX and MGG, we previously proposed efficient procedures for the inference of genetic interactions based on the experimentally observed time-course data of system components (mRNA).
There are many network candidates of gene expression which can realize the same experimentally observed facts, however, the structures of these network candidates of gene expression are different each other. Therefore, we should propose efficient analytical method for extracting useful information from many network candidates of gene expression.
We previously proposed an analyzing procedure for extracting core interactions from many network candidates, and confirmed that the sensitivity of kinetic parameters included in common core binomial interactions is significantly greater than that included in other unique interactions. The interactions with having high sensitivity much contribute to realize experimentally obtained time-course data of gene expression network. We will be able to use these interactions as clue when we investigate about organizationally complex system. It is difficult to use such common core binomial interactions to analyze the dynamic behavior on the gene expression network even if becoming a clue that investigates important relations between genes. Therefore, in this tutorial, we shall describe about the efficient method for extracting common core binomial interactions of an enough number to analyze the dynamic behavior of the gene expression network.

Parameter Estimation Methods in Biochemical Systems Theory: Classical and Evolutionary Approaches

Speakers: Ricardo del Rosario (UPD, Philippines and MPI Biochem, Germany) and Prospero Naval, Jr (UPD, Philippines)


Digital Libraries and Workflow Processes for Systems Biology
Speaker: Su-Shing Chen (U Florida, USA)


We will survey the field of digital libraries and its extension
to workflow processes, which has been contributed by the speaker during the last 10 years and plus. Then we will discuss the relevance of these two topics to systems biology and survey the requirements of building such systems.

Extracting Network Models from Pathway Databases
Baltazar Aguda (Ohio State U, USA)

This tutorial begins with an overview of the nature of hundreds of pathway-related databases now available on the internet, and ends with an example of how a network model was derived by integrating information from these databases. Key methods involved in the extraction of network models include modularization and qualitative network analysis; these will be
illustrated with a modeling problem concerned with the entry into the mammalian cell division cycle - a cellular process that is compromised in most cancers.


1:00-2:00 pm Tuesday

W1: Biological Systems Analysis: Crop and individual plant properties in the context of engineering C4 rice.
John Sheehy A. B. Ferrer, K.G. Tan, and F. Danila (International Rice Research Institute, Philippines)


A system can be defined as a number of interacting elements existing within a boundary which is surrounded by an environment. Biological systems are hierarchical in the sense that they can be described at different levels of detail stretching from molecular to organismal. The success of an operation at any level depends on the successful integration of processes at the lower levels. However, it is important to bear in mind that when the system is viewed as a whole, it is expected that the whole delivers more than the simple sum of its parts. The nature of the emergent properties of the product determine the value added to the inputs and ultimately the efficiency of the system. Progress in understanding the behavior of complex natural systems begins with observations at the whole system level and a top down approach is necessary to guarantee significant achievements from genetic engineering of plants to enhance yields. The success of biological systems depends on control mechanisms that are often ill understood. There needs to be an increased awareness of plasticity: the properties of plants that change when plants are grown as individuals or as members of plant communities. For example, changes in specific leaf area and photosynthesis that occur when individual plants become members of a dense community can be overlooked by those operating at a molecular level. Success in producing a C4 rice plant will depend on targeting the genes controlling several key characteristics of leaves that retain their efficacy, are not plastic, regardless of the plant’s environment.

Rice Genetic Modification for Improved Nutrient Content
Speaker: Rhodora Almedita (PhilRice, Philippines)


The GCP Information System
Speaker: Richard Bruskiewich (International Rice Research Institute, Philippines)


Avner Friedman (Ohio State U, USA)
Keynote Talk Title: Multiscale Models of Tumors
Keynote Talk Abstract:

The multiscale features in the model we consider are both temporal and spatial. Temporal, because in addition to time t we introduce time spent by the cells in one of their phase cycles; spatial, because we work with both cells density (a macroscopic quantity) and gene mutation (a microscopic quantity).

At the restriction point of phase G1, the cell must decide whether to go into the S phase, apoptosis, or the quiescent phase G0. A similar decision is made just before the cell is ready to go into mitosis. The above decisions are affected by the cell’s environmental conditions, e.g., hypoxicity, overpopulation, etc. When some genes are mutated, the decision to go into S may be made in spite of unfavorable environmental conditions, such hypoxic conditions, and this leads to cell proliferation and initiation of tumor.

After formulating a general model, I shall deal with the question of the effect of oxygen concentration on tissue’s health, and state several mathematical results and hypotheses.

Paper Title: Inferring the mechanistic basis for the dynamic response of the MyD88-dependent and –independent pathways
Authors: K. Selvarajoo, M. Helmy, M. Tomita, and M. Tsuchiya
Presentor: Kumar Selvarajoo (Keio University, Japan)

The Toll-like receptors (TLRs), play a critical role in mammalian first line of defense against invading pathogens by recognizing pathogen-associated molecular patterns (PAMPs) such as lipopolysaccarides (LPS). We constructed dynamic in silico model of TLR4 signaling, constituting the Myeloid Differentiation factor 88 (MyD88)-dependent and –independent pathways and investigated the experimental induction of MyD88-dependent pathway genes Tnf, Il1b, Cxcl1 and the MyD88-independent pathway genes Ccl5, Cxcl10, Ifit1 in wildtype and several knock-out (KO) conditions. By fitting our model with wildtype experimental data and analysing with MyD88 KO, TRIF KO and MyD88 and TRIF (double) KO conditions, we infer that the crosstalk between TRIF to TRAF6 and TRIF to TAB/TAK complex via RIP1 is key for the quantitative induction of MyD88-dependent pathway genes Tnf, Il1b, Cxcl1 in MyD88 KO conditions. Our systemic approach provides a novel perspective to dynamic TLR4 pathways analysis.

Paper Title: Patent Prosecution in Systems Biology Software
Authors: M-F Lee and D. Fernandez
Presentor: Ming-Fen Lee (Fernandez & Associates LLP, USA)

This paper presents an overview of intellectual property (IP) rights and systems biology, by introducing the concepts and issues of the IP rights applicable to systems biology, and, particularly, discussing relevant IP implications for systems biology software. Additionally, recent events in the United States, including several Supreme Court’s decisions, the United States Patent and Trademark Office new rules, and the Patent Reform Act of 2007, which might transform the United States patent system, are addressed whenever appropriate. Various topics of technology transfer, such as licensing and business agreements, however, are not covered in depth.

Paper Title: Multi-objective Evolutionary Optimization for Inferring S-System Models of Biological Networks
Authors: P-K Liu and F-S Wang
Presentor: Pang-Kai Liu (National Chung Cheng University, Taiwan)


The inference of biological networks, such as gene regulatory networks, protein-protein interaction networks, metabolic pathway networks etc., from time-course data is one of the main challenges in systems biology. The ultimate goal of inferred modeling is to obtain expressions that quantitatively understand every detail and principle of biological systems. We introduce an interactive inference algorithm to infer a realizable S-system structure for biological networks. The inference problem is formulated as a multiobjective optimization problem to simultaneously minimize the concentration error, slope error and interaction measure in order to find a suitable S-system model structure and its corresponding model parameters. The multiobjective optimization problem is solved by the constraint method to minimize the interaction measure with subject to the expectation constraints for the concentration and slope error criteria. The approach could avoid assigning a suitable penalty weight for sums of magnitude of kinetic orders for the penalty problem in order to prune the model structure.

Poster Title: Extending Biochemical System Theory to Hybrid Modeling by Means of Functional Petri Nets
Authors: J. Wu and E.O. Voit
Presentor: Jialiang Wu (Georgia Tech, USA)

Most biological processes are by nature hybrid systems consisting of interacting discrete and continuous components, which renders it desirable to have a modeling framework which is capable of combining deterministic and stochastic, discrete and continuous effects. In the context of systems biology, such a combination is encountered in the integration of biological pathways (the gene regulatory pathways, metabolite pathways and signaling pathways) with on-off decisions and stochastic components into one comprehensive modeling system. The integration is not trivial because it is well recognized that hybrid systems are difficult to set up, analyze and control.

Biochemical Systems Theory (BST) is a very successful modeling and analysis framework for predicting continuous, deterministic dynamical behaviors of systems in biology and medicine. As in other dynamical approaches, BST is implicitly based on the assumption that relatively large numbers of molecules interact freely within a well-mixed medium. This assumption is important because it permits the use of (continuous) rates and, thus, the use of differential equations. However, many intracellular behaviors are discrete and random in nature, for instance, because of low substrate concentrations or heterogeneous reaction environments, which suggests the need for tools that are able to represent and analyze mixtures of continuous and discrete events.

In this project, we have begun to extend the well-established BST modeling methodology to hybrid modeling within the framework of Hybrid Functional Petri Nets (HFPN). First we implemented GMA and S-system models in a standard Petri Net framework. We then included discrete, stochastic and delay effects. Using small test cases we performed comparative analyses and simulations to access the feasibility, quality and efficiency of the hybrid method.

So far, we have used a toggle switch and small metabolic pathways as test cases for the proposed, combined modeling platform. Our long-term goal is to integrate gene regulatory, signaling, and metabolite networks into comprehensive modeling systems that will eventually allow us to understand system dynamics under normal and various stress conditions.

Paper Title: Approximate kinetic formalisms for modeling metabolic networks: does anything work?
Authors: A. Sorribas, E. Villaprinyo, and R. Alves
Presentor: Albert Sorribas (University of Lleida, Spain)

Approximated kinetic formalisms provide useful mathematical representations for modeling complex genetic and metabolic networks. Although the different formalisms that currently have a wider use for modeling biological systems have technical similarities, their practical utility differs. A formal comparison of the rational behind each formalism help understanding their similarities and differences. From a practical point of view, we discuss basic requirements of a useful mathematical description and provide a rational for selecting a particular formalism depending on the purpose of the study. Such an analysis, suggests that the power-law formalism has clear advantages if one’s purpose is that of analyzing circuit design and systemic properties. The recently proposed Saturable and Cooperative formalism can provide an additional tool that allows for a complementary analysis of the predictions based on the power-law formalism.

Paper Title: Semi-automated Reconstruction of Biological Circuits
Authors: R. Alves, E. Villaprinyo, and A. Sorribas
Presentor: Rui Alves (University of Lleida, Spain)

Large amounts of molecular data regarding most aspects of cellular functioning are accumulating. These data range from sequence and structural data to gene and protein regulation data, including time dependent changes in the concentration of all cellular molecules. Integration of the different datasets through computational methods is required to efficiently organize and extract biological information that is relevant from a Systems Biology perspective.
In this paper we discuss how different computational tools and methods can be made to work together integrating different types of data, mining these data for biological information, and assisting in pathway reconstruction and biological hypothesis generation. We propose an algorithm for the integration of data and discuss an example of its application. This algorithm can then be used to generate testable biological hypothesis, creating an iterative theoretical/experimental loop.

Poster Title: Mathematical Analysis of the Dynamics of a Single-Strain HIV Model with Multiple Endemic States
Authors: L.K. Faina, L. Almocera, and P. Sy,
Presentor: Lindley Kent Faina (University of the Philippines-Visayas)

We analyze a third order autonomous, nonlinear system of ordinary differential equations modeling the dynamics of a single-strain HIV. We establish boundedness of
solutions and positive invariance of a certain region Γ in the nonnegative orthant. We show the existence of the disease free equilibrium and the multiple endemic states for a certain range of parameter values. We also compute the basic reproduction number R0 and show its relation to the stability of the steady states. We carry out numerical
simulations to show the possible bifurcations that the system undergoes.

Paper Title: EUCLIS: Towards a digital library – based Common Information Space for Chronobiologists
Authors: R. Santos and E. Mendoza
Presentor: Roselyn Santos (Ludwig Maximilians University, Germany)

The EUCLOCK Information System (EUCLIS) aims to develop an information infrastructure that will support the cooperative work of both the experimenters and modelers of the worldwide chronobiology community. The current architecture of EUCLIS integrates a database system that supports the storage and analysis of experimental data, and a digital library system of internal and external references. Due to the continuously increasing diversity in the information requirements of the EUCLOCK chronobiologists, a more efficient framework for EUCLIS is a loosely-coupled digital library system architecture that allows more flexibility in adding information into the system and in interacting with 3rd party systems such as collaborative workflow management systems.
The EUCLIS architecture is envisioned to have two main components: a Digital Library as information space (EUCLib) and a Common digital Laboratory for collaboration (EUCLab). The EUCLib will be based on the DELOS DLS Reference Architecture1 and will serve as a space for information like experimental results, genes, references, biomaterials, SOP, classes, images, tools, models, and ontology. The EUCLab will serve as a collaborative workspace that will contain the common lab notebook and workflow management components. This will support the research procedures of experimenters and modelers, and will allow people to work across time and space boundaries.

Paper Title: A Computational Model of Dopamine Metabolism and Parkinson's Disease
Authors: Z. Qi, G.W. Miller, and E.O. Voit
Presentor: Zhen Qi (Georgia Tech, USA)

Alterations in dopamine metabolism have been implicated in a variety of disorders including Parkinson's disease, schizophrenia, and attention deficit hyperactivity disorder. Advancements in the diagnosis and treatment of these disorders have been impeded, in part, by our inability to assess how various genetic, environmental, and metabolic factors impact the underlying biochemical mechanisms of dopamine metabolism. The difficulties in comprehending the consequences of alterations in dopamine metabolism are due to the facts that the pathways involved are complex, tightly regulated, and not amenable to direct measurements in humans. As an aid that might facilitate diagnosis and the development of treatment regimens, we have begun to merge currently available information and expert knowledge regarding dopamine metabolism into a computational model, following the guidelines of biochemical systems theory. After subjecting our model to mathematical diagnosis and analysis, we made direct comparisons between model predictions and experimental observations and found that the model exhibited a high degree of predictive capacity with respect to genetic and pharmacological changes in gene expression or function. Using this model, we screened out key components and processes as well as their combinations, which may be associated with development of PD. While the proposed model of dopamine metabolism is clearly preliminary, future extensions and refinements might eventually serve as an in silico platform for prescreening potential therapeutics, identifying immediate side effects, screening for biomarkers, and assessing the impact of risk factors of disease.

Paper Title: Cephalostatin 1-induced Apoptosis in Leukemic Cells: From Petri Net Model to Kinetic Model
Authors: E. Rodriguez, R. del Rosario, A. Rudy, A. Vollmar, E. Mendoza
Presentor: Eva Rodriguez (University of the Philippines, Diliman)

In this study, we describe a Petri net model for the apoptotic signal transduction pathways induced by cephalostatin 1 on leukemic cells. We validate the Petri net model by structural and invariant analyses. The resulting T-invariants reveal biological interpretations consistent with known behavior of cephalostatin 1-stimulated apoptosis. In addition, we set up a kinetic model corresponding to this Petri net model.

Paper Title: Determination of bottleneck enzymes in a metabolic reaction system by dynamic sensitivity analysis
Authors: F. Shiraishi and Y. Suzuki
Presentor: Fumihide Shiraishi (Kyushu University, Japan)


A method for determination of bottleneck enzymes in a metabolic reaction using dynamic sensitivity values, i.e., dynamic logarithmic gains, has been discussed. A mathematical model for a fed-batch penicillin A fermentation process is employed as an application example. It is found that the order of the magnitudes of the dynamic logarithmic gains changes during the fermentation process, in other words, the candidate of the most likely bottleneck enzyme changes with time. To determine the bottleneck enzyme, therefore, dynamic logarithmic gains are time-averaged over the whole fermentation time and these values are used as an indicator for determination of the bottleneck enzyme. When a metabolic reaction system is in a non-steady state during the whole fermentation period, the order of the magnitudes of the time-averaged dynamic logarithmic gains does not necessarily agree with the order of the final concentrations of a desired metabolite when the activities of bottleneck enzyme candidates are finitely changed. In this case, it is necessary to check not only the order of the magnitudes of the time-averaged dynamic logarithmic gains but also the time courses of the metabolite concentration at the relevant pool. In a steady state or pseudo-steady state, on the other hand, the bottleneck enzyme can directly be determined from the order of the magnitudes of the dynamic logarithmic gains.

Paper Title: Variability dynamics of protein levels in human cells
Authors: A. Sigal, R. Milo, A. Cohen, and U. Alon
Presentor: A. Sigal (Caltech, USA)

We investigated variability and its temporal dynamics in proteins in human cells. We measured fluctuations in endogenous proteins which were fluorescently labeled by introducing YFP as an additional exon into the coding sequence of one genomic allele. Variability (s.d./Mean) ranged between about 0.15 and 0.3 for the proteins studied. Protein level differences between cells were transient: cells higher than average could become lower and vice versa. The timescale for the auto-correlation to decay to half was typically several cell generations. The relatively long auto-correlation time observed in protein levels may account for some of the individuality in cell behavior.

Paper Title: MADMan: A Benchmarking Framework for Parameter Estimation in Biochemical Systems Theory Models
Authors: R. del Rosario, M.T. Echavez, M.T. de Paz, P.C. Zuniga, M.C.R. Bargo, C.O. Talaue, C. Arellano, J.M. Pasia, P.C. Naval, E.O. Voit, E. Mendoza
Presentor: Ricardo del Rosario (University of the Philippines and Max Planck Institute of Biochemistry,Germany)

MADMan is a consortium of institutions in Munich, Atlanta, Diliman (Quezon City, Philippines) and Manila that has the goal to benchmark estimation algorithms used in Biochemical Systems Theory (BST). In recent years, due to the increasing availability of in-vivo time-resolved measurements of biochemical systems, considerable effort has been exerted by different groups to present parameter estimation algorithms to fit BST models to experimental data. The comparison and evaluation of these different published algorithms is a huge task necessitating cooperation between different groups. Moreover, since there is currently no algorithm which can satisfactorily handle all BST parameter estimation problems, the project is expected to take a bit of time to complete.

Poster Title: Location of Malaria Infection in Cells through Grammatical Evolution of Image Processing Techniques (MICE)
Authors: C. Clarin and P. C. Naval
Presentor: Christine Clarin (University of the Philippines-Diliman)

Malaria is one of the most important to address among all the tropical diseases. Over two billion people are at risk of infection and at least a million people die every year cite. Over the years, the most common tool in investigating malaria infection is the blood film examination. This paper proposes a grammatical evolution approach in automatically generating an image analysis JAVA program which effectively detects and locates malaria infection in blood smear images. After detection and location, the system should immediately get the ratio of infected over healthy cells to serve as quantification of parasite density in the blood.

This study is relevant in that it aims to lessen and speed up the labor intensive process of manually analyzing bulk image samples. It should also produce results which are competitive with the efficiency and accuracy of malaria detection operation and reference personnel. Consequently, the JAVA codes from the system should evolve to effectively analyze images subjected under different conditions such as laboratory preparation, image magnification, light orientation etc.

Paper Title: Implementing brane calculus
Authors: M.P. David, A. Phillips, E. Mendoza, and L. Cardelli
Presentor: Maria Pamela David (University of the Philippines-Diliman)

Brane calculi are a family of process calculi where actions are tightly coupled to membranes. In contrast to stochastic pi-calculus, brane calculus provides a framework for spatial organization. Correlating the spatio-temporal location of an object with its capabilities is a very important aspect in modeling biological phenomena that range from biochemical reactions to viral pathogenesis. The primitives offered by brane calculus should provide a more natural environment for modeling events where the occurrence or non-occurrence of a reaction is dependent on where the reactants are, their compatibility, as well as on non-deterministic factors. Currently, however, no tool exists for implementing Brane calculus models. Here, a tool for implementing brane calculus models over the Stochastic Pi Machine (SPiM) is described.

Paper Title: Design of a Homogeneous Ensemble for Splice-site Identification in Human Sequences
Authors: J. Pabico, E. Mojica, and J.R. Micor
Presentor: Jaderick Pabico (University of the Philippines-Los Banos)

This paper outlines a methodology for designing and implementing two novel combination methods for an ensemble of neural network models that classify donor and acceptor sites in human genome sequences. Using a data set of 3,190 sequences, the majority-vote ensemble reduced the false sites recognized by the best stand-alone neural network model by more than 1% while the weighted-vote ensemble reduced it further by more than 3%.

Poster Title: Agent-based models for biochemical systems
Authors: T-Y. Wang, K-C. Chen, D.F. Hsu, C-Y Kao
Presentor: Tse-Yi Wang (National Taiwan University, Taiwan)

Motivation: Mathematical models in molecular systems biology are based on quantitative methods to describe the integrated behavior of complex biochemical networks. Such approaches often use variables and equations to model multiple components of the systems and their evolvements over time. Agent-based models provide another framework for complex biological systems, focusing on how individuals behave and what integrated
behavior emerges. We are motivated to construct agent models for S-systems and explore the relations between them.

Results: To model a complex biochemical network, the
dynamics of integrated behavior is characterized by an S-system as a whole, and a large number of individual entities are considered as autonomous decision-making agents separately. Our goal is to establish the generalized birth-death processes of agents dominated by probabilistic rules such that the integrated influxes and effluxes can be expressed as multivariate power-law
functions. Our agent-based models for biochemical systems are basic and provide an intuitive approach to complex systems more completely, including the process of integrated behavior emerges from individual interactions. It implies that advanced applications
of mathematical models with agent-based modeling might be well suited to portray complex biochemical networks.

Paper Title: A Mathematical Model of Light-Induced ATP Synthesis in the Archaea Halobacterium salinarum
Authors: S. Veflingstad, D. Oesterheld, E.O. Voit
Presentor: Siren Veflingstad (Georgia Tech, USA and Max Planck Institute for Biochem, Germany)

The archaea Halobacterium salinarum can live with light as the only energy source, due to a light-driven proton pump in the membrane. The energy in the resulting proton gradient is utilized by ATP synthase for synthesis of ATP. This system has been studied quite extensively, both in terms of characterizing the major membrane proteins, as well as the accompanying flow of ions and accumulation of ATP. However, seemingly contradictory experimental results regarding the properties of some of the ion transporters and non-intuitive changes in pH show that there are crucial parts of the system that we still do not sufficiently know or understand. Here we present early results from the interdisciplinary development of a mathematical model of light-induced ATP synthesis within the framework of Biochemical Systems Theory.

Paper Title: Gene expression levels in Saccharomyces cerevisiae strongly reflect energetic cost rates of gene products
Authors: E. Villaprinyo, R. Alves, A. Sorribas, and A. Salvador
Presentor: Ester Villaprinyo (University of Lleida, Spain)

We examine how protein abundances under log phase (LP) growth and mRNA change-folds in the shift from LP to late stationary phase (SP) relate to the energetic cost rates (E) of maintaining each polypeptide chain’s steady state abundance. In LP polypeptide chain abundances decrease very significantly with estimated per-molecule energetic cost (e). Turnover numbers, primary sequence length, and mean amino acid biosynthetic cost are significantly lower for abundant vs. rare proteins, but only the former two factors contribute substantially for variation in e. In the LP->SP shift, mRNA change-folds significantly decrease — mainly owing to stronger down-regulation — with increasing polypeptide chain cost rates, abundance and length, but increase with turnover numbers. Estimated relative polypeptide chain abundance change-folds significantly decrease with increasing initial abundance, per-molecule energetic costs, polypeptide chain length and turnover numbers. Altogether, the results highlight that: (a) adaptive plasticity of e, mediated by adaptation of mean aminoacid biosynthetic costs, polypeptide chain length and turnover numbers, dampens the selective pressure on gene expression levels for minimizing E; (b) but nevertheless expression levels are very sensitive to this selective pressure, even for polypeptide chains with relatively low values of E.

Paper Title: Computational Identification of Dynamic Biological Networks: Inverse Modeling Approach and Parameter Estimation Strategies
Authors: I-C Chou, M. Vilela, J.S. Almeida, and E.O. Voit
Presentor: I-Chun Chou (Georgia Tech, USA)

In this presentation we propose a strategy to extract implicit information about biological systems from global dynamic data with methods of parameter estimation and structure identification, using the modeling framework of Biochemical Systems Theory (BST).

Poster Title: Exploring Characteristics of Metabolic Networks Lactococcus lactis IL1403
Authors: I-F Chung, S-Y Yang, S-Y Huang, and F-S Wang
Presentor: I-F Chung (National Yang-Ming University, Taiwan)
n this study, we take lactic acid bacteria (LAB) as an example to explore characteristics of metabolic networks by two processes. Firstly, we adopt inverse flux analysis to obtain the whole internal regulations of LAB metabolic networks. In this way, it is easier to infer the unknown metabolic flux from the known ones and view the flux changes of the integral metabolic networks by interfering with some metabolic flux. Furthermore, we use elementary flux modes (EFMs) combining quadratic programming approach to reconstruct physiological flux distributions of fermented metabolism in LAB, attempting to find out the significant pathways of the main products.

Poster Title: EUCLIS - Towards an Information Infrastructure for the Chronobiology Community
Authors: E. Mendoza, R. Santos, R.T. Batista, M.C. del Rosario, M.P. David, C. Clarin, A. Lao, and R. Bernardo
Presentor: E. Mendoza

The EUClock Information System (EUCLIS) endeavors to provide the network of EUCLOCK chronobiologists a suitable infrastructure for inter-communication, data sharing and secure data storage. EUCLIS aims to support the cooperative work of researchers, both experimental processes (storage of results, methodological annotation, accessibility, appropriate visualization tools) as well as modelling of circadian systems. The architecture of EUCLIS is a hybrid of a database system for experimental data and models, and a digital library system for references. Experimenters and modellers can base their work on common basis of components containing diverse information, including clock genes catalogue, internal and external references, multimedia data and software tools.

Poster Title: Halobacterium salinarum R-1 Metabolism: Reconstruction, Modeling & Analysis
Authors: O. Gonzalez, S. Gronau, M. Falb, F. Pfeiffer, E. Mendoza
Presentor: Orland Gonzalez (Max Planck Institute for Biochem, Germany)

We present a genome-scale metabolic reconstruction for the extreme halophile Halobacterium salinarum. The reconstruction represents a summary of the knowledge regarding the organism's metabolism and has already led to new research directions and improved the existing annotation. We used the network for computational analysis and studied aerobic growth of the organism using dynamic simulations in media with 15 available carbon and energy sources. Simulations resulted in predictions for the internal fluxes which describe at the molecular level how the organism lives and grows. We found numerous indications that cells maximized energy production even at the cost of longer term concerns such as growth prospects. Simulations showed a very low carbon incorporation rate of only ~15%. All of the supplied nutrients were simultaneously degraded, unexpectedly including five which are essential. These initially surprising behaviors are likely adaptations of the organism to its natural environment where growth occurs in blooms. In addition, we also examined specific aspects of metabolism, including how each of the supplied carbon and energy sources is utilized. Finally, we investigated the consequences of the model assumptions and the network structure on the quality of the flux predictions.

Paper Title: On the Structural Complexity of DNA Templates and some Wider Implications
Author: G. Yagil
Presentor: Gad Yagil (Weizmann Institute of Science, Israel)

The concept of biocomplexity is in heavy need for a quantitative evaluation procedure. We have been formulating for some time a procedure suitable for the numerical assessment of molecular and structural complexities, based on the Algorithmic Complexity concept of Kolmogoroff and Chaitin (yagil, 2000). Our procedure is applied here to the genomic material of the cell, namely DNA to evaluate the complexity of selected DNA segments. The DNA segments evaluated include: 1) Two E. coli gene of a protein (lac I and lac Z), as an example of DNA sequences which are as complex as possible (relative complexity = ~1); 2) the telomere of a yeast chromosome, which has a considerable number of regular features, and is indeed specified in a special short RNA code; and 3) a segment of human DNA, gene p53, which has a number of regular features such as 29 interspersed alu elements. These features cause a certain reduction in complexity for the p53 gene, but do not invalidate the overall conclusion that DNA base composition is the dominant factor in determining cellular complexity. The high complexity DNA sequences are a necessary result of the template character of DNA and reflect the role of genomic DNA as a principal control element of a cell. It will be a challenge to find systems of lower complexity, with the ability to respond to challenges from the cell’s environment to the extent that templated systems can do. Cellular complexity and template directed activity are thus highly intertwined properties, at the root of most developmental, behavioral and evolutionary processes.

Paper Title: Noisy: Identification of Problematic Columns in Multiple Sequence Alignents
Authors: A.W.M. Dress, C. Flamm, G. Fritzsch, S. Grunewald, M. Kruspe, S.J. Prohaska and P.F. Stadler
Presentor: Andreas Dress (MPG/CAS Partner Institute for Computational Biology, China)

Sequence-based methods for phylogenetic reconstruction from (nucleic acid) sequence data are notoriously plagued by homoplasies and alignment errors. As most protein-coding genes show dramatic variations in substitution rates that are not uncorrelated across the sequence, this often leads to a patchwork pattern of phylogenetically informative and effectively randomized regions. Alignment errors accumulate in the highly variable regions of an alignment. The program noisy implements a method that identifies effectively randomized sites in a multiple sequence alignment based on assessing the distribution of character states along an (appropriately determined) cyclic ordering of the taxa. Removal of these sites appears to improve the performance of phylogenetic reconstruction algorithms if there are sufficiently many taxa (say, at least a dozen) in the data set.

Poster Title: Parameter Estimation in S-systems using the Newton Flow
Authors: M. de Paz. R. del Rosario, and E. Mendoza
Presentor: Margie de Paz (University of the Philippines-Diliman)

The Newton Flow method for parameter estimation in
S-System models of biochemical systems was presented in [1]
where its successful performance in parameter estimation in a 4-dimensional and 30-dimensional problem, together with its
applicability to network inference, was illustrated. Here, we study the performance of the algorithm to solve parameter estimation problems in other biochemical systems which were not studied in [1]. In this presentation we will concentrate on our results on the simple genetic network which was published in 1996 [2]. The Newton Flow algorithm was designed to solve the optimization problem arising from parameter estimation in the “decoupled” S-system problem [3]. The presence of a 1-dimensional attractor for Newton Flow was discussed in [1] but
the mathematical proof to show its existence in general was not established and seems to be a difficult problem. Thus, it is
necessary to test the applicability of the algorithm for each
application of the S-system model.

Paper Title: Model driven optimization based on standard formalisms
Authors: A. Marin-Sanguino, N.V. Torres, E. Mendoza, and D. Oesterhelt
Presentor: Alberto Marin-Sanguino (Max Planck Institute for Biochem, Germany)

Model driven optimization is an important part of metabolic
engineering and has been the subject of very different ap-
proaches. One of the most common is to use standard mod-
els with an approximate simplified structure. The many
similarities among different methods have often been over-
looked due to their use of completely different notations.
Two such methods, one for s-system models and one for lin-
log, are shown in this work and translated to a common
notation that facilitates comparison among them. Further-
more, these methods are combined into an extended version
that can be used with the Generalized Mass Action formal-

Paper Title: Programming Language Support in Simulating Semantic Multimodal Network Models
Authors: A. Sioson and L. Heath
Presentor: Allan Sioson (Ateneo de Naga University, Philippines)

MMNpl is an interpreted programming language used to
express the tedious and involved computations typical in
calculating the state of biological entities or biots involved in
a biological network model using the semantic multimodal
network (MMN) formalism. Here, we describe the design
of MMNpl and the use of an MMNpl interpreter software,
called mmnsim, in facilitating simulation runs of semantic
MMN models.

Paper Title: Growth and Ligninolytic System Production Dynamics of the Phanerochaete chrysosporium Fungus: A Modelling and Optimization Approach
Authors: J. Hormiga, J. Vera, I. Frias O. Wolkenhauer, and N.V. Torres-Darias
Presentor: J. Hormiga (University of La Laguna, Spain)

The well documented ability to degrade lignin and a variety of complex chemicals showed by the white-rot fungus Phanerochaete chrysosponum (Aitken et al. 1989; Bumpus 1989) has made it the subject of many studies in areas of environmental concern, including pulp bioleaching and bioremediation technologies. However, until now most of the work in this field have been focused on the ligninolytic system but very little have been done to understand the biochemical and regulatory structure that could explain the growth dynamics, substrate utilization and the ligninolytic system production, in spite of the fact that this type of information is required for (or could facilitate) the optimization tasks aiming to improve the process economic feasibility.

In this work we want to fill this gap by applying the principles and approaches of System Biology to this problem. We have investigated the growth dynamics, substrate consumption and lignin peroxidase production of the Ph. chrysosporium wild type (MUCL 19335) under a set of definite culture conditions. Based on data gathered from different authors (Tien and Kirk; 1988; Dosoretz et al, 1993) an in our own experimental analysis we have built a model using a power-law representation in the GMA version that have been used as a platform to make predictive simulations. The model that incorporates some qualitative biological information and experimental data on the kinetic structure of the process incorporates some basic assumptions about the underlying biochemical system (structure and regulation). Provided with we were able to estimate the system’s parameters from a time series experimental measurements by means an algorithm previously adapted and optimised for power-law models (Vera et al., 2006). The model was subsequently checked for its quality through the comparison of its predictions with the experimental behavior observed in new, different experimental settings and through perturbation analysis aimed to test the model robustness. Hence, the model showed to be able to predict the dynamics of two critical variables such as biomass and lignin peroxidase activity when in conditions of nutrient deprivation and after pulses of veratryl alcohol. Moreover, the model successfully predicts the variables evolution during both, the active growth phase and after the deprivation shock. The close agreement between the predicted and observed behavior set up the basis for a further understanding of the kinetic structure and its regulatory features and provide the necessary background for optimization studies based in the system’s mathematical description.

Poster Title: A Biochemical Systems Theory Model of Oxidative Phosphorylation in Halobacterium Salinarum
Authors: C. Talaue, R. del Rosario, E. Mendoza, D. Oesterhelt
Presentor: Cherryl Talaue (University of the Philippines-Diliman)

Halobacterium salinarum is a rod-shaped halophilic archaeon. It can live under four bioenergetic regimes, namely: aerobic respiration, photosynthesis, anaerobic respiration and arginine fermentation. Halobacterium salinarum is considered a model organism for photosynthesis due to its simple mechanism for energizing the membrane: it has a light-driven proton pump bacteriorhodopsin which is the simplest known ion pump. A lot of experimental studies have been performed on photophosphorylation in H. salinarum, but less attention has been devoted on its oxidative phosphorylation. In this work, we present a mathematical model of respiration in H. salinarum using the mathematical modeling framework of Generalized Mass Action (GMA) which one of the variants of Biochemical Systems Theory (BST).

Poster Title: In Silico Dynamical Analysis of Cellular Systems: A Molecular Perturbation Approach
Authors: T.M. Perumal, W. Yan, and R. Gunawan
Presentor: Thanneer Malai Perumal (National University of Singapore)

The complexity of a typical cellular network limits the use of human intuition in understanding how functional regulation is accomplished in a cell. Mathematical modeling and analysis in systems biology offer a quantitative approach in tackling such problem. A novel dynamical analysis based on sensitivities to molecular perturbation is introduced. The result of this analysis can illustrate a dynamical picture on how regulation and/or signalling is accomplished in a given network. The analysis is then applied to a model of cell death regulation in jurkat T-cell line to show the usefulness of this new method.

Poster Title: In Silico Simulation of the Terpenoid Metabolic Network
Authors: A.H Hawari, Z.A.M. Hussein, and N.M. Nor

Presentor: Aliah Hazmah Hawari (Universiti Kebangsaan Malaysia)

Terpenoids are secondary metabolites which are expressed under extreme conditions. The terpenoid network system was chosen as these compounds has great potential in food industries, pharmaceutical as well as organic compound industries. The knowledge or information that could be extracted from the terpenoid biochemical network could be manipulated and applied in genomics and proteomics approaches to increase the production of compounds. This research focuses on the computational approach to design a model that enables an observation on the terpenoid biosynthesis network. The model is based on Ordinary Differential Equations (ODEs) in defining the reactions in the metabolic network and MATLAB (SimBiology) is used to construct and simulate the model.

Poster Title: Identification of the biochemical response variables to glycerol pulse in E. coli by a multivariate approach
Authors: D.V. Guebel, M. Canovas, N.V. Torres-Darias
Presentor: Daniel Guebel (Biotechnology Counseling Services, Argentina)

In a previous communication we presented some evidences (1) that glycerol uptake in E. coli under aerobic-batch culture conditions could occur through a membrane channel. Herein, by using a multivariate approach, we have analyzed the cellular response in terms of twelve biochemical variables after perturbation by a glycerol pulse of a steady, continuous E.coli culture operated in anaerobiosis, with high biomass density (2).

We concluded that in anaerobiosis glycerol is not taken-up by a Michaelian mechanism, but instead, shows a biphasic pattern with a sharp reduction in the influx rate (137 folds) despite the high glycerol external availability. This finding provides additional support to our previous claim that glycerol transport might be subjected to some kind of biochemical control rather than controlled by the instantaneous glycerol availability (i.e., the hysteresis effect).

Moreover, by using partial least squares regression (PLS) and Orthogonal Least Squares (OLS) correction –and in spite that all variables experienced significant variations after the pulse– we have identified which variables are primarily responsive to the glycerol perturbation. The most relevant findings arising from this analysis are: a) E. coli response does not imply a constant responsive structure along the monitored response period (120 min); b) At the early response step (0-5 min), after the extraction of the OLS, we detected 3 independent PLS factors; c) From these, only the first one has a close linear relationship with the glycerol input while the remaining behaved with an oscillatory, glycerol-independent pattern; d) Acetate production cannot be explained by the glycerol pulse, which implies that the overflow metabolism should be discarded as the cause of the acetate production; e) The NADH/NAD ratio showed a dual control: it is mainly glycerol-independent while in a lesser proportion is associated to the glycerol input; f) ATP strongly correlates with the glycerol input as well as ACS, ICL, ICDH enzyme activities and the formate yield. We conclude thus that at early response, ATP is mainly produced at the substrate level rather than from NADH oxidation. Moreover, given that the NADH/NAD ratio remained constant around 0.6, its behavior seems to be dissociated from the ATP dynamics; g) CHR enzyme showed also a dual pattern control since most of its changes are disconnected from the glycerol pulse, but a minor fraction it does; h) Carnitine biosynthesis shows also no correlation with the glycerol consumption rate, nor with the required CHR enzyme; i) Carnitine (which requires ATP for its synthesis) appears as independent of both, glycerol and ATP, whereas CHR has a strong correlation with ATP. Thus, CHR and ATP could not be limiting under the testing conditions; j) During early response, ethanol and lactate dynamics seems to be mutually independent being both also independent of the glycerol perturbation.

The next step in this investigation will be to integrate the obtained evidences in a general picture of E. coli physiology through its mathematical modeling, enabling thus a more rational manipulation of this microorganism.

Poster Title: Avoiding Catabolite Repression using Systems Biology
Authors: A. Sevilla, M. Canovas, C. Gonzalez-Alcon, N.V. Torres-Darias, and J.L. Iborra
Presentor: Angel Sevilla (University of Murcia, Spain)

Signal transduction pathways are usually avoided when optimizing a biotransformation process since they require complex mathematical formulations. The aim of this work was to use a Systems Biology approach to optimize and monitor the biotransformation of L-carnitine using signal transduction pathways. To this end, a dynamic model was constructed, integrating the metabolic pathways of L-carnitine biosynthesis as well as the expression of this metabolism by means of its regulation by transcription factors such as cAMP-CRP and CaiF. The model was validated using different C-sources as well as different reactor feeding approaches. A linear relationship between the external cellular cAMP and the L-carnitine production levels was predicted before being experimentally confirmed in several scenarios. Moreover, results of the model simulations and subsequent experimental findings demonstrated that the addition of exogenous cAMP was able to restore the L-carnitine production when glucose was used as C-source

Poster Title: 13C-NMR to monitor online the kinetics of intracellular metabolite pools in response to heat stress: input data for modeling the trehalose cycle in Saccharomyces cerevisiae
Authors: L. Fonseca, C. Sanchez, J. Wu, H. Santos, and E.O. Voit
Presentor: Luis Fonseca (Universidade Nova de Lisboa, Portugal)

Trehalose is widely distributed in living cells where it plays a variety of roles that are generally associated with protection against stress. The disaccharide is frequently found in yeast, fungi, and plants, but also occurs in many bacteria and hyperthermophilic archaea. In Saccharomyces cerevisiae, the intracellular concentration of trehalose increases rapidly in response to many environmental stresses, including heat stress. The high trehalose levels have been correlated with tolerance to adverse conditions and led to the notion that trehalose functions as a chemical chaperone. The objective of the present work is to understand the design and operation of the trehalose cycle in S. cerevisiae through a combination of experimental and computational approaches. Here, we revisit earlier work, which assumed that trehalose production is transcriptionally controlled. In conflict with this assumption, newer metabolomics data from our lab show that trehalose increases much too fast to be driven by genomic mechanisms. Instead, our preliminary modeling analysis suggests that observed heat induced changes in the activity of key enzymes might be sufficient to evoke the observed responses. The data for this analysis came in part from published data and in part from new real-time metabolic time course studies using in vivo NMR methods. Specifically, a circulatory system was used to pump the yeast cell suspension between a mini-reactor and the NMR tube in a 500 MHz spectrometer. Temperature, pH, and pO2 were controlled in the bioreactor. A pulse of [1-13C]-glucose was added and the time courses of labeled metabolites were monitored under control conditions and also during heat stress (39ºC). Cells accumulated small amounts of trehalose (2-4 mM) under control conditions, while FBP reached 18 mM. Under heat stress (10 min 39ºC), trehalose accumulation reached 8 mM following a pulse of 65 mM glucose. When the duration of the stress was increased to one hour and 3 pulses of glucose were supplied, trehalose increased to 25 mM. The experimental system also allowed monitoring end-product formation (ethanol, glycerol and acetate). The time courses of FBP and trehalose build-up combined with data at the transcriptional and transductional level of relevant genes of the trehalose cycle were used as input data for a series of computational models.

Poster Title: An Ant Colony Optimization Algorithm for Parameter Estimation and Network Inference Problems in S-System Models
Authors: P. Zuniga, J. Pasia, H. Adorna, R. del Rosario, and P.C. Naval
Presentor: Philip Zuniga (University of the Philippines-Diliman)

In this paper, we propose to use Ant Colony Optimization in the parameter estimation of S Systems. We plan to implement two forms of the ACO for this work. The first one is the an ACO for discrete network inference, while the second one is an ACO for solving continuous problems.

Paper Title: Energetic constraints in adaptive gene expression responses of yeast under environmental changes
Authors: E. Villaprinyo, A. Salvador, R. Alves, A. Sorribas
Presentor: E. Villaprinyo (University of Lleida, Spain)

A successful adaptive response of yeast under stress requires the synthesis of protective molecules that help minimizing cellular damage. The amount and type of these molecules depends on the type of stress. We search for functional constraints that can explain fine tuning of gene expression under stress. For instance, under resource depletion, one may expect to find downregulation of the expression of large and abundant proteins, and upregulation of the expression of shorter proteins. Such a trend may be less evident if the stress does not compromise resource availability. In this work we analyze the existing data and find evidence that is consistent with economy in metabolism as an important pressure for shaping regulation of proteins synthesis in yeast stress response.

Poster Title: Comparisons of the performances of three methods for dynamic sensitivity computation
Authors: F. Shiraishi, K. Hattori, and H. Hirayama
Presentor: F. Shiraishi (Kyushu University, Japan)


The calculation of dynamic logarithmic gains consists of derivation of differential equations for sensitivities by partially differentiating differential equations for metabolite concentrations (Procedure 1) and computation of dynamic sensitivity values by simultaneously solving the two differential equations (Procedure 2). The following three methods have been compared: Method 1 (Procedure 1 is performed by numerical differentiation and Procedure 2 is by the Taylor series method), Method 2 (Procedure 1 is performed by analytical differentiation and Procedure 2 is by the Taylor series method) and Method 3 (Procedure 1 is performed by analytical differentiation and Procedure 2 is by an implicit method). The accuracy of calculation is found to be higher in the order of Method 2, 1, and 3, while the calculation time is shorter in the order of Method 3, 2, and 1. On the other hand, Methods 2 and 3 require commercially available expensive software. Thus, it is recommended that these three methods should be selected according to the requirements.

Poster Title: Heptylprodigiosin Induces Apoptosis in Various Carcinoma Cell Lines and Triggers Death Receptor Pathway in Jurkat T Cell Leukemia
Authors: G. Ranches, A. Rudy, G. Concepcion, and A. Vollmar
Presentor: Glory Ranches (University of the Philippines-Diliman)

In this paper, we report the apoptosis-inducing effect and mechanism of action of Heptylprodigiosin (HPDG), a marine microbial metabolite, in various carcinoma cell lines and Jurkat T cell leukemia, respectively. We show that HPDG induces DNA fragmentation in MCF-7 breast cancer cells, SK-OV-3 ovarian cancer cells and Jurkat T cells, and chromatin condensation, as well. We also demonstrate that HPDG-induced apoptosis is caspase-dependent and Fas receptor-mediated in which caspase-8 and FADD are required to activate downstream cascade of the anti-Fas-derived signal in Jurkat T cells, and that PARP cleavage is elicited through the activation of caspase-3/caspase-9 resulting in DNA fragmentation. Our findings indicate that HPDG could be a potent sensitizer in tumor cells with Fas overexpression and a lead compound in chemotherapy-based combinatorial regimen.

Poster Title: Regulation of the FNR System
Authors: D. Tolla and M. Savageau
Presentor: Dean Tolla (University of California-Davis, USA)

The long-term goal of this work is to characterize the relationship between cycling and physiology of the FNR network of Escherichia coli. In this report, we formulate a model of this system, fit its parameters to existing experimental data, and predict results that were not used in the fitting. The results show excellent agreement between predictions and experimental data.

Paper Title: Global Nonlinear Root-Finding using the Canonical S-system Form
Authors: R. Fasani and M. Savageau
Presentor: Rick Fasani (University of California-Davis, USA)

Many problems in computational geometry, engineering, and applied mathematics--including finding the steady state of a system of ordinary differential equations (ODEs)--ultimately reduce to a search for roots of one or more equations. Previous work suggested that an S-system methodology produced an efficient nonlinear root-finding algorithm. The method was promising, but did not guarantee that all roots would be found in every case. Our recent work has led to a better understanding of the problem, and has created a foundation for at least one novel nonlinear root-finding algorithm. We show that a set of nonlinear equations can be recast to an S-system, and show that finding the steady-state solution, or the roots of the system, reduces to finding the intersections of concave surfaces. We then present a basic divide-and-conquer algorithm that utilizes axis-aligned bounding boxes. We provide a simple example, and show that the method can be extended to the general case.

Paper Title: Global Tolerance of Biochemical Systems and the Design of Moiety-Transfer Cycles
Authors: P. Coelho, A. Salvador, M. Savageau
Presentor: P. Coelho (University of California-Davis, USA)

Analyses of biochemical systems tend to emphasize local aspects of performance, i.e. systemic responses to small changes about the nominal values of concentrations and parameters. However, important indices of metabolic performance that remain almost constant in the neighborhood of the operating point may abruptly break down at some distance, often leading to pathological consequences. Currently, there is no generic approach to identifying and characterizing the boundaries where the local performance of a biological system deteriorates abruptly. Here we introduce a generic approach to the characterization of boundaries between operational regimes based on the piecewise power-law representation of the relevant rate laws. This conceptual framework allows us to precisely define and quantify “global tolerance” as the ratio between the normal value of a parameter and the value at such a boundary. We illustrate the utility of these concepts in the context of the moiety-transfer cycle, which is a form of coupling between pairs of reactions that is very prevalent in metabolism. Our results show that the region of “best” local performance is surrounded by “poor” regions, and that selection for improved local performance often pushes the operating point away from regime boundaries, thus increasing global tolerance. These results suggest that selection for effective system design may lead to large “safety factors” that protect the system from excursions into regions of poor local performance. These predictions are found to be in agreement with experimental data from the NADPH redox cycle of human erythrocytes.

For inquiries, contact:

Dr. Rafael P. Saldaña
Member, Organizing Committee, ICMSB '08
Member, Program Committee, ICMSB '08
Chair, Workshop on Systems Biology of Rice
Tel. +63 2 4266125, +63 2 7090907
Mobile: 0928-5043121
E-mail: raffysaldana@gmail.com

No comments: