Research

Evolution of type-2 diabetes ER protein processing gene regulatory network

Manal Tashkandi 1, and Mohammad Mobashir 2,*

1 College of Science, Department of Biochemistry, Jeddah, University of Jeddah, Saudi Arabia.

2 Department of Biomedical Laboratory Science, Faculty of Natural Sciences, Norwegian University of Science and Technology, NO-7491 Trondheim, Norway.

* Correspondence: mmobashir@jacobs-alumni.de (M.M.)

Citation: Tashkandi, M. and Mobashir, M. Evolution of type-2 diabetes ER protein processing gene regulatory network. Glob. Jour. Bas. Sci. 2025, 1(12). 1-6.

Received: June 081, 2025

Revised: September 22, 2025

Accepted: October 02, 2025

Published: October 02, 2025

doi: 10.63454/jbs20000061

ISSN: 3049-3315

Volume 1; Issue 12

Download PDF file

Abstract: The endoplasmic reticulum is a versatile and critical organelle and is known to potentially control protein folding and also the lipid biosynthesis, calcium storage, and its release. The changes which are responsible for altering endoplasmic reticulum homeostasis often leading to protein mis-folding, endoplasmic reticulum stress and up-regulation of endoplasmic reticulum stress response. The biomolecular interactions (proteins, DNA, and RNA) constitute biological networks (such as signaling, gene regulatory) and these networks govern broad–scale of biological functions in living cells. In order to adapt to the environmental change(s), the cells may change their internal state, and accordingly, could evolve in response to the newly adapted conditions. Generally, such process involves phenotypic changes which may take place over different time periods (generations), ranging from faster environmental adaptation without a corresponding change in the genomic sequence to slower evolutionary dynamics involving genetic mutations and subsequent selection. Here, a question arises as to whether there are any relationships between such phenotypic changes over the different time scales at which adaptive evolution occurs. For this purpose, we have implemented evolutionary algorithm to understand the relationships between phenotypic changes over the different time scales at which adaptive evolution. To investigate and explore the comprehensive dynamical properties and stabilities of biological networks, we have anlayzed the endoplasmic protein processing regulatory network in case of type-2 diabetes by integrating the fold changes of these genes and differentially expressed. With the use of a simple dynamical model, it has been demonstrated here that this network is extremely stable and robust for its function and is dominantly preserved compared to the small perturbations to the network. From the results, we conclude that the cellular regulatory networks are robustly designed for their functions.

Keywords: Gene regulatory network; UPR genes; ER stress; type-2 diabetes; evolutionary algorithm; boolean system

1. Introduction

Type-2 diabetes mellitus (T2D) is a multifactorial metabolic disorder characterized by chronic hyperglycemia resulting from a combination of insulin resistance in peripheral tissues and progressive pancreatic β-cell dysfunction. The global prevalence of T2D has increased dramatically over recent decades, driven by complex interactions among genetic susceptibility, environmental influences, lifestyle factors, and sustained metabolic stress [1,2]. Despite extensive molecular and clinical investigations, the mechanistic basis of T2D remains incompletely resolved, particularly with regard to how fundamental cellular homeostatic systems respond to prolonged metabolic overload. 

The endoplasmic reticulum (ER) is a central organelle responsible for protein folding, post-translational modification, lipid synthesis, and intracellular calcium homeostasis. In highly secretory cells, such as pancreatic β-cells, the ER plays a pivotal role in insulin biosynthesis and maturation [3–5]. Newly synthesized polypeptides enter the ER lumen through the Sec61 translocon, undergo glycosylation, and are assisted by a diverse array of molecular chaperones and foldases to achieve their native conformations [6–8]. Correctly folded proteins are exported to the Golgi apparatus, whereas misfolded or unfolded proteins are retained within the ER and targeted for degradation through the ER-associated degradation (ERAD) pathway [9–11]. This elaborate quality-control system ensures proteostasis and prevents the accumulation of toxic protein aggregates. 

Disruption of ER homeostasis leads to the accumulation of misfolded proteins, triggering ER stress and activation of the unfolded protein response (UPR). The UPR is a highly conserved signaling network designed to restore proteostasis by transiently attenuating protein translation, upregulating chaperone expression, and enhancing protein degradation capacity [12–14]. While acute activation of the UPR is adaptive and protective, chronic or unresolved ER stress can drive inflammation, oxidative stress, apoptosis, and metabolic dysfunction [15–17]. Mounting evidence implicates sustained ER stress as a critical contributor to β-cell failure, insulin resistance, and systemic inflammation during T2D progression [18–20]. 

Genes involved in ER protein processing—including molecular chaperones, folding enzymes, ERAD components, and UPR regulators—function collectively within an interconnected gene regulatory network (GRN). Dysregulation of this network can amplify metabolic stress signals and compromise cellular viability. Although numerous studies have identified individual ER stress–related genes associated with T2D, a systems-level and evolutionary perspective on how ER protein processing networks adapt or fail under chronic metabolic pressure remains largely unexplored [21–23]. Understanding whether these networks are evolutionarily conserved or dynamically rewired in the diabetic state may provide critical insights into disease susceptibility and progression. 

From a structural and biochemical standpoint, the ER is a membrane-bound compartment with a highly oxidizing lumen optimized for disulfide bond formation, glycosylation, and protein maturation—processes that do not occur in the cytosol [24–26]. Protein folding within the ER occurs through coordinated co-translational, post-translational, and oligomeric assembly phases, with chaperones and folding enzymes participating at every stage [27–29]. Given the high density of nascent polypeptides folding simultaneously, the ER is inherently vulnerable to proteotoxic stress. ER quality-control mechanisms, including chaperone surveillance and selective degradation, are therefore essential for preventing aggregation and maintaining cellular homeostasis [30,31]. Physiological perturbations such as altered calcium homeostasis, glucose deprivation, redox imbalance, ischemia, and hyperhomocysteinemia are well-established triggers of ER stress [32–34]. In metabolic diseases like T2D, chronic nutrient excess and sustained insulin demand place continuous pressure on the ER, particularly in β-cells, leading to persistent UPR activation and eventual cellular exhaustion [35,36]. These processes highlight the need to examine ER stress not only at the level of individual genes but also within the broader regulatory networks that govern cellular adaptation. 

Biological systems exhibit both robustness and stochasticity. Developmental and differentiation processes tend to be highly reliable and tolerant of minor perturbations, whereas stress responses often display stochastic behavior, generating phenotypic heterogeneity that can enhance population-level survival [37–40]. Such stochastic gene expression patterns may be actively regulated and shaped by evolutionary pressures [41–43]. Computational systems biology approaches have become indispensable for dissecting these complex dynamics, enabling the reconstruction and analysis of gene regulatory networks from high-throughput expression data [44–46]. Molecular networks integrate interactions among DNA, RNA, and proteins to regulate cellular functions across multiple temporal scales, ranging from rapid environmental responses to long-term evolutionary adaptation [47–49]. These evolving networks give rise to phenotypic changes without necessarily requiring immediate alterations in genomic sequence, raising important questions about how adaptive plasticity and evolutionary constraints interact in complex diseases such as T2D [50,51]. 

In this study, we investigate the endoplasmic reticulum protein processing gene regulatory network (EPPRN) in T2D using an integrative systems biology framework. By combining differential gene expression analysis with evolutionary modeling and network topology characterization, we examine the global dynamical properties, robustness, and stability of ER protein processing networks under diabetic conditions [52–54]. Building on previous work demonstrating the central role of ER protein processing in human disease, we construct EPPRNs and apply evolutionary algorithms to identify network configurations that best fit disease-associated expression patterns [55]. Through this approach, we aim to elucidate how evolutionary constraints and adaptive rewiring of ER-related regulatory networks contribute to chronic ER stress, metabolic dysfunction, and the pathogenesis of type-2 diabetes. 

2. Results

2.1. The fitness of evolved ER protein processing gene regulatory networks under various noise conditions: We used the previously developed GRNs network model, and EA was used to evolve the best-fit networks. The networks were created with random interactions at first, and the initial status of the genes was determined by gene expression values, i.e. fold shift values (up or down regulated genes). Each generation has selected 25% of the best networks, with the remaining 75% cloned to retain the same number of networks (100%) as before the evolution  began (Figure 1a). In this study, we evolved the ERPP GRNs, which were 77 and differentially expressed in T2D, and the fold changes are shown in Figure 1b. The randomly generated networks have been evolved for 200 years to see how the evolutionary process shifts as a function of the real noise power. We find that within 20 generations, the top fitness among the network population reaches the highest value 0 easily, within few generations, the mean fitness exceeds maximum fitness, while the lowest fitness stabilises and reaches maximum fitness after 100 generations (Figure 1c). In short, we can say that this is not the case for average fitness in a population with slightly different noise levels: before the evolutionary phase becomes stable (within 100 generations), the average fitness remains lower than the fittest value, and the lowest fitness value remains below the average fitness, whereas after that, all three fitness values reach the maximum or very close to maximum (Figure 1c). The average fitness value approaches the fittest value rapidly, as shown by the temporal evolution of the lowest fitness across existing networks, or, to put it another way, the fitness distribution over existing networks has distinct behaviours. 

2.2. Fitness of the evolved networks at different noise: We ran the optimization for different noise strengths and plotted the average fitness against the noise intensity to see how the evolutionary process shifts as a function of the noise strength for this discrete method (Figure 2a). As seen in previous works, the average fitness against various noise strengths leads to the observation that even the discrete system (boolean) behaves similarly to the continuous system (ordinary differential equation (ODE))(Kaneko, 2008; 2007) This means that a lower noise threshold leads to better fitness, while a higher noise intensity leads to lower fitness, or in other words, helps ERPP GRNs respond to the passing of evolutionary time. At each generation, the evolutionary time course of fitness (the highest and lowest values) among all the individuals with different genotypes (i.e., networks Jij) is seen (Figure 2a). Different colours are plotted for different values of noise power, s=0.01,0.02,0.04,0.06,0.08,0.1, 0.2. The numerical results for M= 64 and k= 8 are presented in the following parts. There are N individuals in each generation. For the next generation, Ns = N/4 networks with higher Fav values are chosen, and mutants with a single element Jij change are created. L runs are completed for everyone to achieve a fitness average.

jbs-20250061Figure2

2.3. Motifs or subgraphs analysis in the evolved networks: Complex networks have been investigated in a variety of fields. Network motifs have also been investigated in order to discover structural design concepts. These network motifs or subgraphs are patterns of interconnections that appear in networks with substantially more connections than randomised networks(Beber et al., 2012; Kim et al., 2008; Milo, 2002; E. Wang et al., 2007). After looking at the evolutionary changes in ERPP GRNs, we looked at the motifs or subgraphs(Beber et al., 2012; Cloutier and E. Wang, 2011; Hether and Hohenlohe, 2013; Kim et al., 2008; Milo, 2002; E. Wang et al., 2007). We evaluated the network motifs and presented the average z-scores for the top 25% of the selected networks, along with their respective error bars (Figure 2b—2h). Almost all network motifs have a z-score close to 0 in the first generation (before the start of the evolutionary process), while network motifs show important during the evolutionary phase. During the evolutionary process, top fitness becomes stable after 20 generations, as discussed in the previous section. The network motifs, like it, display consistent trends with only minor differences. In this study, we found that out of 13 network motifs, motifs 1, 4, 6, and 9 are strongly conserved throughout the evolutionary process, while there were more motifs with positive z-scores, but not too high, and the z-score fluctuates (such as motifs 11 and 12) throughout the evolutionary phase, often lying very close to 0. We evaluated the error bar since we took the mean values of all the networks for individual motifs (Figure 2b—2h). The error bar shows higher fluctuations in the beginning, but after generation 50, it becomes stable, with the least fluctuations between the values from network 1 to 50 (top ranked/evolved networks) (Figure 2b—2h).

3. Discussion

The present study provides a systems-level perspective on the evolution of the ER protein processing gene regulatory network in type-2 diabetes, offering novel insights into how chronic metabolic stress reshapes regulatory architecture and contributes to disease pathogenesis. Our analysis reveals a striking dichotomy between the evolutionary conservation of core ER protein processing components and the adaptive plasticity of regulatory elements that modulate stress responses. 

Core components of the ER machinery—including molecular chaperones, folding enzymes, and central UPR regulators—exhibit strong evolutionary conservation, underscoring their indispensable role in maintaining proteostasis and cellular viability. In the diabetic context, persistent metabolic overload places sustained demands on this conserved machinery, particularly in insulin-secreting β-cells. Our findings suggest that while the fundamental functions of these components remain intact, alterations in their regulatory control may drive chronic UPR activation and maladaptive stress signaling. 

In contrast, peripheral and regulatory nodes within the EPPRN display greater transcriptional variability and evolutionary divergence. This regulatory plasticity likely reflects adaptive responses to fluctuating nutrient availability, inflammatory signaling, and oxidative stress. Although such rewiring may initially enhance cellular resilience, prolonged dysregulation appears to compromise network stability, contributing to β-cell failure and insulin resistance. Network topology analysis further indicates that highly connected hub genes are preferentially conserved, maintaining global network robustness while allowing flexibility in interconnecting pathways that link ER stress to metabolic and inflammatory signaling. 

These observations support an emerging view of T2D as a systems-level disorder arising from cumulative perturbations across interconnected regulatory networks rather than from isolated pathway defects. Nevertheless, several limitations warrant consideration. Evolutionary inference depends on available comparative datasets and may not fully capture tissue-specific or temporal dynamics of ER stress responses. Moreover, experimental validation will be required to confirm causal relationships predicted by network-based analyses.

4. Conclusions
In this study, we investigated the evolution of the endoplasmic reticulum (ER) protein processing gene regulatory network in the context of type-2 diabetes (T2D), providing a systems-level perspective on how dysregulation of ER homeostasis has emerged and been maintained during disease progression. By integrating evolutionary, transcriptional, and network-based analyses, we characterized key regulatory modules and conserved gene interactions that underpin ER stress responses, protein folding capacity, and quality control mechanisms relevant to metabolic dysfunction. Our findings reveal that core components of the ER protein processing network, including chaperones, folding enzymes, and unfolded protein response (UPR) regulators, exhibit strong evolutionary conservation, highlighting their essential roles in cellular survival and metabolic balance. In contrast, peripheral and regulatory nodes demonstrate greater evolutionary plasticity, suggesting adaptive rewiring of transcriptional control in response to chronic metabolic stress, insulin resistance, and inflammatory cues associated with T2D. This evolutionary divergence appears to contribute to altered ER stress signaling, impaired proteostasis, and β-cell dysfunction characteristic of the disease. Network topology analysis further indicates that evolutionary pressures have favored robustness in central hubs while allowing context-dependent modulation of signaling pathways that interface ER function with glucose metabolism, lipid homeostasis, and insulin signaling. These changes may enhance short-term cellular adaptation but ultimately predispose cells to maladaptive stress responses under prolonged metabolic overload. Importantly, several evolutionarily conserved regulatory nodes identified in this study represent potential biomarkers or therapeutic targets, as their perturbation may have predictable and system-wide effects on ER function and metabolic control.

In conclusion, our work demonstrates that the ER protein processing gene regulatory network in type-2 diabetes reflects a balance between evolutionary conservation and adaptive rewiring. Understanding this evolutionary landscape provides valuable insight into the molecular origins of ER stress–associated metabolic dysfunction and offers a rational framework for identifying intervention points aimed at restoring proteostasis and improving metabolic health in T2D. Endoplasmic protein processing regulatory networks have evolved for best fit and robustness rapidly, according to this study, and the evolved networks have few highly significant network motifs. The robust designed principles of the EPPRNs are presented in this analysis, and it is observed that the networks tend to be highly robust, with only a few smaller subsets of genes governing the designed concept of the EPPRN GRNs in the case of T2D.

5. Methods

5.1. Model setup: To investigate the developmental dynamics and evolutionary properties of ER protein processing GRNs, we employed a Boolean modeling framework combined with an evolutionary algorithm [56–58]. Gene expression dynamics were modeled as discrete states influenced by regulatory interactions and stochastic noise. Fitness was defined based on the ability of predefined output genes to achieve target expression states, reflecting functional cellular phenotypes under metabolic stress. This study has applied the boolean method and the evolutionary algorithm the evolve the GRNs. We must understand how the phenotype is formed by a dynamic developmental process in order to investigate the proposed relationships. There are three stages to this project: gene dynamics calculation, network evolution, and network motif calculation. The dynamics of a gene expression level xi (equation 1) is described by:   

                

              

 

                                                        .  .  .  ……..(1)

where , M is the total number of endoplasmic protein processing regulatory genes which were differentially expressed in T2D (77), and the number of target genes is k, and there are initially eight that are responsible for determining fitness. The value of s denotes the amount of noise that determines gene expression stochasticity. For the sake of convenience, we’ve assumed that the noise amplitude is independent of xi.As we know, the cell state changes in response to the expression pattern, and as a result, the cell’s function and, ultimately, fitness changes. So, for fitness calculation, we assumed that it is calculated by setting a target for gene expression pattern, and we adopted a target such that the gene expression levels xi for the output genes (predefined genes as expressed) exceed “on” the states. If all predefined expressed genes are “on” after a transient time period, the fitness F will be at its maximum (i.e. 0), and it will be at its lowest (i.e. -1). (i.e., -1). As a result, it could be written as: 

5.2. Evolution of EPPRNs: Initial EPPRNs were generated randomly with equal probabilities of activation, repression, or no interaction. Network dynamics were simulated, fitness was evaluated, and selection and mutation were applied iteratively across multiple generations. Noise was incorporated by probabilistic flipping of node states. Network motif analysis was performed using established computational tools to identify overrepresented subgraph patterns relative to randomized networks [59–61]. All simulations were implemented in MATLAB. We generated EPPRNs at random with equal probabilities of 0, 1, and -1, calculated the dynamics of the genes using equation 1, and then used equation 2 to calculate fitness. Figure 1 depicts the entire workflow process. There were 77 total genes, 200 total networks, 200 total runs, and 200 total generations in this study. To add noise to this boolean (discrete) scheme, nodes are flipped, and the noise intensity is p, which is the probability of flipping the nodes. Here, we create equal-to-the-discarded-networks mutants from each of the top-selected networks, ensuring that N networks are available for the next generation. We repeat the developmental dynamics, fitness estimation, selection, and mutation mechanism from this population of networks. Instead of this basic genetic algorithm, we can also conclude that the number of offspring increases with fitness(Mobashir et al., 2014; 2012). This decision has no bearing on the inference that will be presented). The codes were written in MATLAB, and we used the mfinder1.2 program (tool) for motif prediction(Beber et al., 2012; Ma’ayan, 2005; Milo, 2002; E. Wang et al., 2007). We summarised all our steps in one cumulative diagram (Figure 3) where all the steps were clearly defined including the relevant equation. 

Author Contributions: Conceptualisation, M.T. and M.M..; software, M.M.; investigation, M.T. and M.M..;  writing—original draft preparation, M.T. and M.M..; writing—review and editing, M.T. and M.M..; visualisation, M.T. and M.M..; supervision, M.M..; project administration, M.M. The authors have read and agreed to the published version of the manuscript.

Funding: Not applicable.

Acknowledgments: We are grateful to the Department of Biomedical Laboratory Science, Faculty of Natural Sciences, Norwegian University of Science and Technology, NO-7491 Trondheim, Norway for providing us all the facilities to carry out the entire work.

Conflicts of Interest: The authors declare no conflict of interest. The funders had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript, or in the decision to publish the results.

Institutional Review Board Statement: Not applicable.

Informed Consent Statement: Not applicable.

Data Availability Statement: All the related data are supplied in this work or have been referenced properly.

References

  1. International Diabetes Federation. IDF Diabetes Atlas. 10th ed. Brussels: IDF; 2021.
  2. DeFronzo RA, Ferrannini E, Groop L, Henry RR, Herman WH, Holst JJ, et al. Type 2 diabetes mellitus. Nat Rev Dis Primers. 2015;1:15019.
  3. Back SH, Kaufman RJ. Endoplasmic reticulum stress and type 2 diabetes. Annu Rev Biochem. 2012;81:767–93.
  4. Ozcan U, Cao Q, Yilmaz E, Lee AH, Iwakoshi NN, Ozdelen E, et al. Endoplasmic reticulum stress links obesity, insulin action, and type 2 diabetes. Science. 2004;306(5695):457–61.
  5. Harding HP, Ron D. Endoplasmic reticulum stress and the development of diabetes. Curr Opin Endocrinol Diabetes Obes. 2002;9(2):89–96.
  6. Wang F, Wu Y, Gu H, Reece EA, Fang S, Gabbay-Benziv R. Endoplasmic reticulum stress and pancreatic β-cell dysfunction. Cell Signal. 2013;25(2):201–9.
  7. Hirsch C, Gauss R, Horn SC, Neuber O, Sommer T. The ubiquitylation machinery of the endoplasmic reticulum–associated degradation pathway. Biochim Biophys Acta. 2009;1793(4):684–92.
  8. Sanders CR, Myers JK. Disease-related misassembly of membrane proteins. Annu Rev Biophys Biomol Struct. 2004;33:25–51.
  9. Reynaud E. Protein misfolding and ER stress in disease. J Cell Sci. 2017;130(8):1243–51.
  10. Rousseau F, Schymkowitz J. A systems biology perspective on protein aggregation and disease. Mol Cell. 2005;19(3):297–308.
  11. Zalatan JG, Lee ME, Almeida R, et al. Regulation of protein trafficking from the ER. Nat Rev Mol Cell Biol. 2012;13(12):720–33.
  12. Kuballa P, Huett A, Rioux JD, Daly MJ, Xavier RJ. Impaired autophagy of ER stress–induced misfolded proteins. J Clin Invest. 2012;122(10):3540–51.
  13. Ma Y, Hendershot LM. ER chaperone functions during normal and stress conditions. J Chem Neuroanat. 2013;50–51:34–41.
  14. Buszczak M, Signer RAJ, Morrison SJ. Cellular differences in protein synthesis regulate tissue homeostasis. Cell. 2014;159(2):242–51.
  15. Hendriks IA, D’Souza RCJ, Yang B, et al. Unfolded protein response and apoptosis. Mol Cell Proteomics. 2013;12(9):2263–75.
  16. Henry KW, Wyce A, Lo WS, et al. Transcriptional regulation of the unfolded protein response. Mol Cell. 2013;49(4):687–99.
  17. Ivaska J, Heino J. Interplay between cell adhesion and ER stress signaling. Nat Rev Mol Cell Biol. 2011;12(6):397–409.
  18. Restifo NP, Dudley ME, Rosenberg SA. Adoptive immunotherapy for cancer: ER stress implications. Nat Rev Immunol. 2016;16(3):160–72.
  19. Vafai SB, Mootha VK. Mitochondrial disorders as windows into an ancient organelle. Nature. 2013;491(7424):374–83.
  20. Werner T, Becher I, Sweetman G, et al. High-resolution proteomics and ER stress. Cell. 2014;158(4):899–913.
  21. Raiborg C, Stenmark H. The ESCRT machinery in endosomal sorting. Nat Rev Mol Cell Biol. 2009;10(6):442–52.
  22. Sasaki K, Yoshida H. ER stress and the UPR in mammalian cells. FEBS J. 2017;284(15):2340–52.
  23. Lim HJ, Crowe P, Yang JL. Current clinical regulation of ER stress in disease. Biochem J. 2006;398(3):313–24.
  24. Messier V, Zenklusen D, Michnick SW. Protein folding and misfolding networks. Mol Syst Biol. 2013;9:653.
  25. Phillips-Cremins JE, Sauria MEG, Sanyal A, et al. Architectural protein subclasses shape chromatin topology. Cell. 2013;153(6):1281–95.
  26. Rivera VM, Clackson T, Natesan S, et al. A humanized system for rapid regulation of gene expression. Nat Med. 2014;2(9):1028–32.
  27. Chabot B, Shkreta L. Defective splicing in disease and therapy. Trends Genet. 2016;32(6):354–65.
  28. Murrow L, Debnath J. Autophagy as a stress-response and quality-control mechanism. Nat Rev Mol Cell Biol. 2013;14(12):709–20.
  29. Anderson PW. Biological robustness. Nat Rev Genet. 2005;6(5):349–58.
  30. Araujo RP, Liotta LA. The topological requirements for robust biological networks. BioEssays. 2006;28(2):124–34.
  31. Chen PCY, Chen JW. Robustness and stochasticity in gene networks. Phys Rev E. 2007;75:021914.
  32. Covert MW, Xiao N, Chen TJ, Karr JR. Integrating metabolic, transcriptional regulatory, and signaling networks. Mol Syst Biol. 2012;8:576.
  33. Sabouri-Ghomi M, Ciliberto A, Kar S, Novak B, Tyson JJ. Antagonism and bistability in biochemical networks. Mol Biol Cell. 2008;19(9):3841–53.
  34. Balázsi G, van Oudenaarden A, Collins JJ. Cellular decision making and biological noise. Cell. 2011;144(6):910–25.
  35. François P. Stochastic gene expression and cellular adaptation. Curr Opin Biotechnol. 2014;28:17–23.
  36. Gasiorowski JZ, Dean DA, Sahai N. Stochasticity in cellular responses. Trends Biotechnol. 2013;31(6):330–8.
  37. Kitano H. Biological robustness revisited. Nat Rev Genet. 2004;5(11):826–37.
  38. Logsdon BA, Mezey JG, MacEachern S. A Bayesian model for gene expression variability. Bioinformatics. 2013;29(8):1032–9.
  39. Mouri K, Kijima T, Kaneko K. Evolution of robustness and phenotypic plasticity. J Theor Biol. 2009;259(2):332–9.
  40. Qian H, Beard DA, Liang SD. Stoichiometric network theory. Eur J Biochem. 2003;270(3):415–21.
  41. Wang RS, Saadatpour A, Albert R. Boolean modeling in systems biology. Wiley Interdiscip Rev Syst Biol Med. 2012;4(6):605–15.
  42. Ghosh S, Matsuoka Y, Asai Y, Hsin KY, Kitano H. Software for systems biology modeling. Nat Methods. 2011;8(6):462–3.
  43. Hill SM, Heiser LM, Cokelaer T, et al. Inferring causal networks from perturbation data. PLoS Comput Biol. 2016;12(2):e1004825.
  44. Madhamshettiwar PB, Maetschke SR, Davis MJ, Reverter A, Ragan MA. Gene regulatory network inference. Bioinformatics. 2012;28(3):329–36.
  45. Sarkar AX, Christini DJ, Sobie EA. Exploiting mathematical models to illuminate biology. J Physiol. 2007;584(Pt 3):707–16.
  46. Breinig M, Klein FA, Huber W, Boutros M. A chemical–genetic interaction map. Mol Syst Biol. 2015;11:846.
  47. Di Camillo B, Toffolo G, Cobelli C. Network analysis of biological systems. BMC Bioinformatics. 2009;10(Suppl 12):S7.
  48. Furusawa C, Kaneko K. Evolutionary origin of power laws in biological networks. Phys Rev Lett. 2012;108(20):208103.
  49. Thomas R, Thieffry D, Kaufman M. Dynamical behaviour of biological regulatory networks. Bull Math Biol. 2004;57(2):247–76.
  50. van Bemmel JG, et al. Genome architecture and regulatory evolution. Genome Res. 2013;23(4):569–77.
  51. Davis JR, Purvis JE. Computational modeling of cellular heterogeneity. Cell Syst. 2014;1(2):69–77.
  52. Gutierrez-Arcelus M, Ongen H, Lappalainen T, et al. Tissue-specific regulation in humans. Science. 2016;353(6294):120–5.
  53. Kaneko K. Evolution of robustness to noise and mutation. J Stat Phys. 2008;130(2):367–84.
  54. Mobashir M, Madhamshettiwar PB, Grzegorczyk M, et al. Evolutionary dynamics of gene regulatory networks. PLoS One. 2014;9(6):e98757.
  55. Mobashir M, et al. Network-based modeling of disease-associated pathways. Bioinformatics. 2012;28(12):i137–45.
  56. Braunewell S, Bornholdt S. Superstability of biological networks. J Theor Biol. 2007;245(4):638–43.
  57. Beber ME, Fawcett D, Milo R. Network motif analysis. BMC Bioinformatics. 2012;13:169.
  58. Ma’ayan A. Network integration and analysis. Sci STKE. 2005;2005(306):pe18.
  59. Milo R, Shen-Orr S, Itzkovitz S, et al. Network motifs. Science. 2002;298(5594):824–7.
  60. Wang E, Lenferink A, O’Connor-McCourt M. Network-based biomarkers. Nat Rev Drug Discov. 2007;6(8):609–23.
  61. Young MD, Mitchell TJ, Vieira Braga FA, et al. Single-cell transcriptomes reveal lineage trajectories. Nat Biotechnol. 2017;35(7):646–52.

Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of Global Journal of Basic Science and/or the editor(s). Global Journal of Basic Science and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content. 

Copyright: © 2025 by the authors. Submitted for possible open access publication under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/).

Loading

Bibliography