Open access

Using Cellular Automata and Global Sensitivity Analysis to Study the Regulatory Network of the L-Arabinose Operon

Written By

Advait A. Apte, Stephen S. Fong and Ryan S. Senger

Published: 08 May 2013

DOI: 10.5772/52085

From the Edited Volume

Emerging Applications of Cellular Automata

Edited by Alejandro Salcido

Chapter metrics overview

1,929 Chapter Downloads

View Full Metrics

1. Introduction

The field of computational biologyhas grown significantly in recent years, allowing researchers to investigatecomplex biological systems in silico. In this chapter, a new method of combining cellular automata (CA) with global sensitivity analysis (GSA) is introduced. This method has the potential to determine which mechanisms of a regulated biological network contribute to characteristics such as stability and responsiveness. For dynamic models of biochemical reactions and networks, determining the correct values ofthe kineticparameters that govern the system is often problematic[12, 13, 19, 29, 30, 33, 36, 46, 47, 49, 50, 55, 57]. This isoften due to the difficulty ofobtaining accurate experimental measurements ofvital kineticconstants[6, 10, 17, 21, 32, 37]. This is currently the case for biochemical reactions occurring in complex environments in vivo that cannot be approximated through more simple experimentsin vitro.Complex gene regulatory networks are a prime example.Modelling these systems using aCA approach allows researchers to easily change kineticparameter values (individuallyand in combination) to study their effects on system function[4, 7, 8, 25, 38, 51]. Since the overall system function (i.e., the output molecule or action) is something that can be measured easily, CA modelling provides a method for determining “difficult-to-measure” parameters using “easy-to-measure” observations of the system being studied.However, the real challenge after conducting a CA study with combinatorial parameter variation is interpreting the results. It has been found that GSAis an extremely useful tool for identifying the parameters that most significantly affect overall model performance[9, 11, 16, 24, 28, 31, 44, 45, 56]. In this chapter,a new approach that combinesCA and GSA is applied for analysing regulatory mechanisms of the L-arabinose (ara) operon. In particular, the influence of the negative autoregulatory (NAR) action of the transcription factor AraC on system dynamics and stability is calculated. The purpose of this chapter is to provide instruction through a detailed example of how to apply CA and GSA simultaneously to analysea biological network that must be studied in vivo. An in-depth explanation of GSA and the regulatory elements of the ara operon are provided in the Introduction. Detailed descriptions of CA model building and system parameters as well as a comprehensive GSA tutorial are presented in the Materials and Methods section.Computational experiments illuminating the influence of the NAR mechanism on the studied regulatory network are presented and discussed throughout the rest of thechapter.

1.1. Global sensitivity analysis

GSA uses Monte Carlo simulations to calculate the outputs of a model over the entire range of all input parameters [39, 40]. This variance-based method calculates the contribution of each input parameter to the total variability of the model output. In other words, GSA is used to determine which inputs most significantly influence the output. This provides the investigator one or multiple targets that can be manipulated to effectively engineer the system. GSA differs significantly from the traditionally used method of partial gradient-based sensitivity analysis (SA). With traditional SA, the change in a model output is calculated by allowing only one parameter to vary, while keeping others constant. This variance in the model output is likely to change if all other parameters are held constant at different values. The GSA approach takes this into account and enables the consideration of multiple parameters simultaneously over the entire range of each parameter. To consider the model output variance caused by only a single parameter is a “first-order” analysis. Two parameters may be considered simultaneously to develop a measure of their interactions in a “second-order” analysis. Or, a single parameter can be considered with is interactions with all other parameters in the model. This is called the “total effect index” [39, 40]. Thus, another significant advantage of GSA over SA is that GSA accounts for the influence and interactions between input parameters over the entire input space. A simplified tutorial of GSA has been published for a deterministic ammonia emissions model [35]. The GSA methods are also presented in detail in this chapter for the ara operon model system.

1.2. The L-arabinose operon

Transcriptional regulation networks are largely made up of recurring regulatory patterns called network motifs. These network motifs have been shown to carry out many signal transduction functions[2, 3, 27, 48]. One of the most abundantly found network motifs is thenegative autoregulation (NAR) motif. In an NAR motif, a transcription factor (TF) negatively regulates the promoter of its own gene or operon. This has been found to

  1. dramatically increase response acceleration and

  2. increase the stability of the gene product concentration response to noise [26, 41].

The L-arabinose system is an example of an NAR network.L-arabinose,which is a five-carbon sugar foundin plant cell walls, is used as a carbon source by many organisms. The ara operon contains genes encoding enzymes leading to L-arabinose catabolism. The selective usage of the ara operon isone of the best-studied gene regulation systems and is well-characterized[5, 15, 22, 23, 34, 42, 52, 59]. The entire arabinose system consists of the following:

  1. the system specific transcription factor (TF) araC,

  2. the arabinose transporters (araE, araFGH, and araJ), and

  3. the ara operon containing the arabinose catalytic enzymes, araBAD.

Ultimately, the ara operon is responsible for the conversion of arabinose to D-xylulose-5-phosphate, which then enters the pentose phosphate pathway. The AraC TF regulates the ara operon [1, 14, 18, 20, 26, 52, 54, 59].Recent studies have shown that araCand the ara operon share a common regulatory protein,cAMP Receptor Protein (CRP),which is activated by cAMP. AraC both activates and represses the ara operonusing a DNA looping mechanism. As a negative regulator, AraC isde-activated by L-arabinose, allowing transcription of the ara operon. AraC represses its own promoter through a NAR motif[1, 14, 26, 43, 58].

1.3. Goals of the Modelling Effort

In this research, the overall influence of NAR on the dynamics of a regulated biological network was studied by applying a unique combination of CA and GSA. Here, the expression of araBAD was calculated in the

  1. presence and

  2. absence of NAR by AraC.

The CA approach was used to simulate this network given altered kinetic rate constants and initial concentrations. Then, the GSA approach was applied to determine which of these parameters most directly influence araBAD expression. When applied to the

  1. presence and

  2. absence of NAR, the difference in GSA results give clues to the influence of the NAR mechanism in regulating the system dynamics.

In the case studied in this research, NAR was found to equally distribute model sensitivity across all input parameters. This dramatically increases stability and responsivenessof the regulatory network. The approach presented in this chapter of combining CA and GSA can be applied to virtually any biologicalnetwork using the methods presented in this chapter.


2. Methods

2.1. Model Construction

The ara operon model was constructed using NetLogo simulation environment [53]. To perform a CA simulation, individual (agents) (i.e., interacting molecules) were allowed to move among (cells) (i.e., spatial locations) inside the simulation environment and undergobiochemical reactions with other agents in their Von Neumann neighbourhood. In all simulations, a two-dimensional16 x 16 matrix of cells was used as the simulation environment, and 10 time steps were executed. Whether a reaction occurs between interacting agents is governed by probability. The agents of the ara operon model are:

  1. L-arabinose,

  2. cAMP,

  3. AraC (the TF regulator),

  4. CRP, and

  5. AraBAD (representing gene products of the ara operon).

Characteristics and reaction rules for individual agents were predefined at model initialization in NetLogo. Basal expression levels for CRP, AraC, and AraBAD were set at 10, 0, and 0 cells respectively.The number of agents occupying cells represents concentration in agent based modelling. For example, CRP was present in 10 cells of the simulation environment upon initializing the simulation. Specifics of the varied model parameters are discussed in detail in the next section. These included

  1. the concentrations of L-arabinose and cAMP,

  2. the probabilities of biochemical reactions, and

  3. NAR by ArgC.

Monte Carlo methods were used to select 2000 values of each parameter to perform CA simulations. This was followed by 1000 independent iterations of the model to perform GSA. The CA simulation records activation of the ara operon measured as number of AraBAD agents present in cells at the end of the simulation. Simulations are also often run to record the number of model “events” required to reach a specified concentration of an agent of interest (e.g., AraBAD). Independent simulations use different values of the varied parameters, resulting in different values of the targeted agent. Two common approaches use

  1. a set number of model events to derive a target agent or

  2. a different number of model events required to reach a specified concentration of the target agent.

The first approach was used in this study. Two different scenarios for NAR byAraC regulation were simulated in this research:

  1. AraC is not allowed to negatively autoregulate its own promoter and

  2. NAR by AraC is allowed.

These simulations seek to understand the influence of the NAR mechanism on overall rigidity and robustness of the ara operon regulatory network.

2.2. System parameters

A mathematical model was created to simulate the dynamics of ara operon activation in the presence and absence of NAR by AraC. CA was applied by allowing critical parameters to vary over thousands of simulations of the system. GSA was then applied to the results to determine which system parameters most influence ara operon activation.The following parameters were taken into consideration while building and simulating the model. The upper and lower bounds of the parameters and a description of their functions are described in detail below.

Simulation 1 2 1000
conc_cAMP 105 188 32
conc_arabinose 159 179 244
rate_CRP_activate 0.019 0.103 0.500
rate_araC_activate 0.418 0.391 0.406
rate_araBAD_activate 0.577 0.326 0.445
rate_araC_autoreg 0 0 0
Output ( Y 1 ) 9 3 ... 296

Table 1.

The M1 matrix of GSA.

  1. conc_arabinose: The initial concentration of L-arabinose was allowed to vary between 1 and 250 cells. Upon binding,L-arabinoseactivates the TF and autoregulatorAraC.Unbound AraC inhibits transcription of the ara operon.

  2. conc_cAMP: Initial concentration of the second messenger cAMP was allowed to vary between 1 to 250cells. cAMP binds to CRP causing its activation. The cAMP-CRP positively regulates transcription of the ara operon.

  3. rate_CRP_activate: This parameter controls the probability of CRP activation by cAMP. This parameter was varied between 0 and 1 for the simulations described in this chapter. This probability parameter represents the rate of activation of CRP.

  4. rate_araC_activate: This parameter controls the probability of AraC activation by L-arabinose and was allowed to vary between 0 and 1.This probabilityultimately controlsthe activity of the araCgene and transcription of the ara operon.

  5. rate_araBAD_activate: This parameter controls the probability of araBAD activation by CRP and was allowed tovary between 0 and 1.

  6. rate_araC_autoreg: This parameter controls the NARbyAraC protein and was allowed tovary between 0 and 1.Thisrepresents the rate at which AraC supresses its promoter.

2.3. Global sensitivity analysis

GSA was performed on the ara operon activation model described in this chapter. The following step-by-step tutorial is given to combine CA with GSA.

The procedure starts with the derivation of the M1 and M2 matrices shown in Tables 1 and 2, respectively. To build each table, 1000 CA simulations were run given random values of the system parameters. For each simulation, the model output (activated araBADor expressed AraBAD) was calculated and recorded. The estimated unconditional means ( E ^ Y ) and estimated unconditional variances ( V ^ Y ) of the model outputs were calculated for both matrices according to the following, where N is the number of simulations (i.e., 1000 for this study).

E Y 1 = 1 N i = 1 N Y 1 ( i ) E Y 2 = 1 N i = 1 N Y 2 ( i ) E1
V Y 1 = 1 N 1 i = 1 N ( Y 1 ( i ) ) 2 ( E Y 1 ) 2 V Y 2 = 1 N 1 i = 1 N ( Y 2 ( i ) ) 2 ( E Y 2 ) 2 E2

Next, the P matrix was created for the calculation of the first-order sensitivity index for each model parameter. To illustrate this example, the model parameter conc_cAMPwas used. The P matrix for this case is shown in Table 3.

The P matrix consists of the conc_cAMPparameter values from matrix M2, and all other parameters are assigned their values from M1. Then model outputs were calculated for the P matrix using these new inputs. Thus, 1000 more simulations are required for each parameter a first-order sensitivity index is desired. The first-order sensitivity index ( S c o n c _ c A M P ) was calculated by the following.

U P = 1 N 1 i = 1 N Y 1 ( i ) Y P ( i ) E3
S c o n c _ c A M P = 1 ( U p E Y 1 E Y 2 V Y 1 ) E4

Next, the total effect index was calculated by creation of the R matrix for each parameter. The first-order index describes the influence of a single parameter on the model output directly. The total effect index takes into account all interactions of a parameter with all other parameters when determining the effect on model output.

Simulation 1 2 1000
conc_cAMP 190 189 227
conc_arabinose 67 212 208
rate_CRP_activate 0.887 0.281 0.671
rate_araC_activate 0.638 0.413 0.693
rate_araBAD_activate 0.530 0.620 0.891
rate_araC_autoreg 0 0 0
Output ( Y 2 ) 608 93 ... 3043

Table 2.

The M2 matrix of GSA.

Simulation 1 2 1000
conc_cAMP (M2) 190 189 227
conc_arabinose (M1) 159 179 244
rate_CRP_activate (M1) 0.019 0.103 0.500
rate_araC_activate (M1) 0.418 0.391 0.406
rate_araBAD_activate (M1) 0.577 0.326 0.445
rate_araC_autoreg (M1) 0 0 0
Output ( Y P ) 12 196 ... 6

Table 3.

The P matrix of GSA.

The R matrix is shown in Table 4 for the conc_cAMPparameter example. To build the R matrix, the parameter values from M1 for conc_cAMPwere used along with parameter values from M2 for all other parameters. An additional 1000 CA simulations are required to calculate the model outputs for the R matrix. The calculation of the total effect index for rcAMP ( S T ( c o n c _ c A M P ) ) was calculated as follows.

U R = 1 N 1 i = 1 N Y 1 ( i ) Y R ( i ) E5
S T ( c o n c _ c A M P ) = 1 ( U R E Y 1 2 V Y 1 ) E6

3. Results

The CA modeling of the ara operon was performed for two cases

  1. without NAR (i.e., negative autoregulation) by AraC and

  2. with NAR by AraC (as is observed experimentally).

GSA was applied to both cases in order to determine how the NAR mechanism impacts overall system dynamics. To simulate the model without NAR, the parameter rate_araC_autoreg was held constant at 0. The results of the GSA calculations derived from Eqs. 1-6 and the values in Tables 1-4 are given in Table 5. This case was simulated without NAR by AraC.

Simulation 1 2 1000
conc_cAMP (M1) 105 188 32
conc_arabinose (M2) 67 212 208
rate_CRP_activate (M2) 0.887 0.281 0.671
rate_araC_activate (M2) 0.638 0.413 0.693
rate_araBAD_activate (M2) 0.530 0.620 0.891
rate_araC_autoreg (M2) 0 0 0
Output ( Y R ) 288 519 ... 324

Table 4.

The R matrix of GSA.

The first-order sensitivity indices for each system parameter for the ara operonmodel without NAR by AraCare reported in Fig. 1. These values are reported as a percentage of the summation of all first-order index values. When interactions between single parameters are not taken into consideration, the probability of CRP activation was found to be the single most important parameter significantly influencing the araBADactivation(29.58%).All other parameters show similar influence (~17%). The total effect indices for the ara operon model without NAR by AraC is shown in Fig. 2. When all the interactions between all parameters were considered, probability of CRP activation (rate_CRP_activate parameter) was shown to have most influence (29.47%) over araBADactivation. The more noticeable result is the small contribution from the initial concentration of L-arabinose (conc_arabinose parameter). This is significant because the ara operon is known to require the presence of L-arabinose to be active in the cell.

Calculation Value
E ^ Y 1 1 (estimated unconditional mean of M1) 256.78
E ^ Y 1 2 (estimated unconditional mean of M2) 228.49
V ^ Y 1 (estimated unconditional variance of M1) 309411.56
V ^ Y 2 (estimated unconditional variance of M2) N/A
U P 4210.45
S c o n c _ c A M P (first-order sensitivity index of conc_cAMP) 0.18
U R 83451.88
S T ( c o n c _ c A M P ) (total effect index of conc_cAMP) 0.94

Table 5.

GSA calculations for the rcAMP parameter

Figure 1.

First-order indices calculated by GSA for the case without NAR by AraC.

The first-order indices for each parameter for the ara operonmodel with NAR by AraC are reported in Fig. 3. By activating the NAR role of AraC, the first-order indices show very close index values for all parameters (~16.5%). In other words, the NAR reduced the excessive influence of CRP activation over araBADactivation. The total effect indices are shown in Fig. 4. A pattern similar to that revealed by first-order indices was obtained. All total effect indices were also similar for all parameters (~16.5%).Adding the NAR by AraCto the regulation network dramatically increased the influence of L-arabinose concentration on araBADactivation.

Figure 2.

Total effect indices calculated by GSA for the case without NAR by AraC.

Figure 3.

First-order indices calculated by GSA for the case with NAR by AraC.


4. Discussion

In this study, a unique combination of CA and GSA were used to study the parameters that influence the dynamics of the ara operon regulatory network. The results of the GSA study revealed the degree to which individual parameters affect the output of a biological model.GSA was used to explorethe influence of NAR on the regulatory network by calculating the impact of parameter variance on model output.Comparing first-order and total effect sensitivity indices with and without NAR by AraC elucidates the roles NAR plays in the signaling network. These include

  1. increasing network stability and

  2. increasing the response of the network to L-arabinose concentrations.

Equal distribution of variation among all parameters suggests that the NAR mechanism increases network robustness,providing protection against random perturbations (both biological and environmental) of the system.

Figure 4.

Total effect indices calculated by GSA for the case with NAR by AraC.

GSA has shown that parameter sensitivity indices can provide useful insight in interpreting the results of CA simulations.Thus, the combination of CA and GSA provides a valuable tool for the identification of source of output variability. In the case of the araoperon, all model parameters showed to contribute equally to the variance of araBADactivationlevel. While all the systemparameters are important and can significantly influence araBADactivation, those parameters with higher first-order sensitivities can have profound effects on regulation of the ara operon if NAR function by AraC is lost. This scenario demonstrates the potential of CA and GSA for identifying targets for manipulating highly interconnected gene regulatory networks.


  1. 1. Akel E. Metz B. Seiboth B. Kubicek C. P. 2009 Molecular regulation of arabinan and L-arabinose metabolism in Hypocrea jecorina (Trichoderma reesei). Eukaryotic cell. 8 1837 44
  2. 2. Alon N. Dao P. Hajirasouliha I. Hormozdiari F. Sahinalp S. C. 2008 Biomolecular network motif counting and discovery by color coding Bioinformatics. 24 i241 9
  3. 3. Apte A. Cain J. Bonchev D. Fong S. 2008a Cellular automata simulation of topological effects on the dynamics of feed-forward motifs. Journal of Biological Engineering 2 2
  4. 4. Apte A. A. Cain J. W. Bonchev D. G. Fong S. S. 2008b Cellular automata simulation of topological effects on the dynamics of feed-forward motifs. Journal of Biological Engineering 2 2
  5. 5. Beverin S. Sheppard D. E. Park S. S. 1971 D-Fucose as a gratuitous inducer of the L-arabinose operon in strains of Escherichia coli B-r mutant in gene araC. Journal of Bacteriology 107 79 86
  6. 6. Califano A. 2011 Striking a balance between feasible and realistic biological models. Science translational medicine 3 103ps39
  7. 7. Chavoya A. Andalon-Garcia I. R. Lopez-Martin C. Meda-Campana M. E. 2010 Use of evolved artificial regulatory networks to simulate 3D cell differentiation Bio Systems 102 41 8
  8. 8. Chavoya A. Duthen Y. 2008 A cell pattern generation model based on an extended artificial regulatory network Bio Systems. 94 95 101
  9. 9. Chhatre S. Francis R. Newcombe A. R. Zhou Y. Titchener-Hooker N. King J. Keshavarz-Moore E. 2008 Global Sensitivity Analysis for the determination of parameter importance in bio-manufacturing processes Biotechnology and applied biochemistry 51 79 90
  10. 10. Chou I. C. Martens H. Voit E. O. 2006 Parameter estimation in biochemical systems models with alternating regression Theoretical biology & medical modelling. 3 25
  11. 11. Ciavatta S. Lovato T. Ratto M. Pastres R. 2009 Global uncertainty and sensitivity analysis of a food-web bioaccumulation model Environmental toxicology and chemistry SETAC. 28 718 32
  12. 12. Coelho F. C. Codeco C. T. Gomes M. G. 2011 A Bayesian framework for parameter estimation in dynamical models PLoS ONE 6 e19616
  13. 13. Cuenod C. A. Favetto B. Genon-Catalot V. Rozenholc Y. Samson A. 2011 Parameter estimation and change-point detection from Dynamic Contrast Enhanced MRI data using stochastic differential equations Mathematical biosciences 233 68 76
  14. 14. Desai T. A. Rao C. V. 2010 Regulation of arabinose and xylose metabolism in Escherichia coli Applied and Environmental Microbiology 76 1524 32
  15. 15. Englesberg E. Squires C. Meronk F. Jr 1969 The L-arabinose operon in Escherichia coli B-r: a genetic demonstration of two functional states of the product of a regulator gene Proceedings of the National Academy of Sciences of the United States of America 62 1100 7
  16. 16. Feng X. J. Hooshangi S. Chen D. Li G. Weiss R. Rabitz H. 2004 Optimizing genetic circuits by global sensitivity analysis. Biophysical journal. 87 2195 202
  17. 17. Frangi A. F. Coatrieux J. L. Peng G. C. D’Argenio D. Z. Marmarelis V. Z. Michailova A. 2011 Editorial: Special issue on multiscale modeling and analysis in computational biology and medicine--part-1. IEEE transactions on bio-medical engineering 58 2936 42
  18. 18. Greenblatt J. Schleif R. 1971 Arabinose C protein: regulation of the arabinose operon in vitro. Nature: New biology 233 166 70
  19. 19. Hadeler K. P. 2011 Parameter identification in epidemic models Mathematical biosciences 229 185 9
  20. 20. Hendrickson W. Schleif R. F. 1984 Regulation of the Escherichia coli L-arabinose operon studied by gel electrophoresis DNA binding assay. Journal of molecular biology 178 611 28
  21. 21. Hinkelmann F. Murrugarra D. Jarrah A. S. Laubenbacher R. 2011 A mathematical framework for agent based models of complex biological networks Bulletin of mathematical biology. 73 1583 602
  22. 22. Irr J. Englesberg E. 1971 Control of expression of the L-arabinose operon in temperature-sensitive mutants of gene araC in Escherichia coli B-r. Journal of Bacteriology 105 136 41
  23. 23. Kolodrubetz D. Schleif R. 1981 Regulation of the L-arabinose transport operons in Escherichia coli. Journal of molecular biology 151 215 27
  24. 24. Li G. Rabitz H. Yelvington P. E. Oluwole O. O. Bacon F. Kolb C. E. Schoendorf J. 2010 Global sensitivity analysis for systems with independent and/or correlated inputs The journal of physical chemistry. A114 6022 32
  25. 25. Machado D. Costa R. S. Rocha M. Ferreira E. C. Tidor B. Rocha I. 2011 Modeling formalisms in Systems Biology AMB Express 1 45
  26. 26. Madar D. Dekel E. Bren A. Alon U. 2011 Negative auto-regulation increases the input dynamic-range of the arabinose system of Escherichia coli BMC Systems Biology 5 111
  27. 27. Mangan S. Alon U. 2003 Structure and function of the feed-forward loop network motif Proceedings of the National Academy of Sciences of the United States of America 100 11980 5
  28. 28. Marino S. Hogue I. B. Ray C. J. Kirschner D. E. 2008 A methodology for performing global uncertainty and sensitivity analysis in systems biology Journal of theoretical biology 254 178 96
  29. 29. Mente C. Prade I. Brusch L. Breier G. Deutsch A. 2011 Parameter estimation with a novel gradient-based optimization method for biological lattice-gas cellular automaton models Journal of mathematical biology 63 173 200
  30. 30. Miro A. Pozo C. Guillen-Gosalbez G. Egea J. A. Jimenez L. 2012 Deterministic global optimization algorithm based on outer approximation for the parameter estima- tion of nonlinear dynamic biological systems BMC bioinformatics 13 90
  31. 31. Mishra S. Deeds N. Ruskauff G. 2009 Global sensitivity analysis techniques for probabilistic ground water modeling MishraS.DeedsN.RuskauffG. 2009. Global sensitivity analysis techniques for probabilistic ground water modeling. Ground water. 47 730 47
  32. 32. Mogilner A. Odde D. 2011 Modeling cellular processes in 3D Trends in cell biology 21 692 700
  33. 33. Neftci E. O. Toth B. Indiveri G. Abarbanel H. D. 2012 Dynamic State and Parameter Estimation Applied to Neuromorphic Systems Neural computation
  34. 34. Ogden S. Haggerty D. Stoner C. M. Kolodrubetz D. Schleif R. 1980 The Escherichia coli L-arabinose operon: binding sites of the regulatory proteins and a mechanism of positive and negative regulation Proceedings of the National Academy of Sciences of the United States of America 77 3346 50
  35. 35. Ogejo J. A. Senger R. S. Zhang R. H. 2010 Global sensitivity analysis of a process-based model for ammonia emissions from manure storage and treatment structures Atmospheric Environment 44 3621 3629
  36. 36. Olufsen M. S. Ottesen J. T. 2012 A practical approach to parameter estimation applied to model predicting heart rate regulation Journal of mathematical biology
  37. 37. Qu Z. Garfinkel A. Weiss J. N. Nivala M. 2011 Multi-scale modeling in biology: how to bridge the gaps between scales? Progress in biophysics and molecular biology. 107 21 31
  38. 38. Rosenfeld S. 2011 Critical self-organized self-sustained oscillations in large regulatory networks: towards understanding the gene expression initiation. Gene regulation and systems biology 5 27 40
  39. 39. Saltelli A. Chan K. Scott M. 2000 Sensitivity Analysis. Wiley.
  40. 40. Saltelli A. Ratto M. Andres T. Camplongo F. Cariboni J. Gatelli D. Saisana M. Tarantola S. 2008 Global Sensitivity Analysis: The Primer. Wiley.
  41. 41. Sarkar R. R. Maithreye R. Sinha S. 2011 Design of regulation and dynamics in simple biochemical pathways Journal of mathematical biology 63 283 307
  42. 42. Schimz K. L. Lessmann D. Kurz G. 1974 Proceedings: Aspects of regulation of the pathways for D-glucose 6-phosphate, D-galactose, and L-arabinose in Pseudomonas fluorescens. Hoppe-Seyler’s Zeitschrift fur physiologische Chemie. 355 1249
  43. 43. Schleif R. 2010 AraC protein, regulation of the l-arabinose operon in Escheric hia coli, and the light switch mechanism of AraC action. FEMS microbiology reviews. 34 779 96
  44. 44. Sheridan R. P. 2008 Alternative global goodness metrics and sensitivity analysis: heuristics to check the robustness of conclusions from studies comparing virtual screening methods Journal of Chemical Information and Modeling 48 426 33
  45. 45. Sin G. Gernaey K. V. Neumann M. B. van Loosdrecht M. C. Gujer W. 2011 Global sensitivity analysis in wastewater treatment plant model applications: prioritizing sources of uncertainty Water research 45 639 51
  46. 46. Sumner T. Shephard E. Bogle I. D. 2012 A methodology for global-sensitivity analysis of time-dependent outputs in systems biology modelling Journal of the Royal Society, Interface / the Royal Society.
  47. 47. Tashkova K. Korosec P. Silc J. Todorovski L. Dzeroski S. 2011 Parameter estimation with bio-inspired meta-heuristic optimization: modeling the dynamics of endocytosis. BMC Systems Biology
  48. 48. Taylor R. J. Siegel A. F. Galitski T. 2007 Network motif analysis of a multi-mode genetic-interaction network Genome biology 8 R160
  49. 49. Tian L. P. Liu L. Wu F. X. 2010 Parameter estimation method for improper fractional models and its application to molecular biological systems. Conference proceedings :... Annual International Conference of the IEEE Engineering in Medicine and Biology Society. IEEE Engineering in Medicine and Biology Society. Conference 2010 1477 80
  50. 50. Toni T. Stumpf M. P. 2010 Parameter inference and model selection in signaling pathway models. Methods in molecular biology. 673 283 95
  51. 51. Tyson J. J. Mackey M. C. 2001 Molecular, metabolic, and genetic control: An introduction. Chaos. 11 81 83 .
  52. 52. Wilcox G. Meuris P. Bass R. Englesberg E. 1974 Regulation of the L-arabinose operon BAD in vitro. The Journal of biological chemistry 249 2946 52
  53. 53. Wilensky U. Netlogo 1999 Center for Connected Learning and Computer-Based Modeling. Northwestern University, Evanston, IL
  54. 54. Winfield M. D. Latifi T. Groisman E. A. 2005 Transcriptional regulation of the 4-amino-4-deoxy-L-arabinose biosynthetic genes in Yersinia pestis. The Journal of biological chemistry 280 14765 72
  55. 55. Yang X. Dent J. E. Nardini C. 2012 An S-System Parameter Estimation Method (SPEM) for biological networks Journal of computational biology a journal of computational molecular cell biology. 19 175 87
  56. 56. Yoon J. Deisboeck T. S. 2009 Investigating differential dynamics of the MAPK signaling cascade using a multi-parametric global sensitivity analysis. PLoS ONE 4 e4560
  57. 57. Zhan C. Yeung L. F. 2011 Parameter estimation in systems biology models using spline approximation BMC Systems Biology 5 14
  58. 58. Zhang L. Leyn S. A. Gu Y. Jiang W. Rodionov D. A. Yang C. 2012 Ribulokinase and transcriptional regulation of arabinose metabolism in Clostridium acetobutylicum Journal of Bacteriology 194 1055 64
  59. 59. Zubay G. Gielow L. Englesberg E. 1971 Cell-free studies on the regulation of the arabinose operon. Nature: New biology 233 164 5

Written By

Advait A. Apte, Stephen S. Fong and Ryan S. Senger

Published: 08 May 2013