%0 Journal Article %1 Dayarian2009Shape %A Dayarian, Adel %A Chaves, Madalena %A Sontag, Eduardo D. %A Sengupta, Anirvan M. %D 2009 %I Public Library of Science %J PLOS Computational Biology %K robustness %N 1 %P 1-12 %R 10.1371/journal.pcbi.1000256 %T Shape, Size, and Robustness: Feasible Regions in the Parameter Space of Biochemical Networks %U https://doi.org/10.1371/journal.pcbi.1000256 %V 5 %X Author Summary Developing models with a large number of parameters for describing the dynamics of a biochemical network is a common exercise today. The dependence of predictions of such a network model on the choice of parameters is important to understand for two reasons. For the purpose of fitting biological data and making predictions, we need to know which combinations of parameters are strongly constrained by observations and also which combinations seriously affect a particular prediction. In addition, we expect naturally evolved networks to be somewhat robust to parameter changes. If the functioning of the network requires fine-tuning in many parameters, then mutations causing changes in regulatory interactions could quickly make the network dysfunctional. For predictions involving gene products being ON or OFF, we found a method that facilitates the study parameter dependence. As an example, we analyzed several competing models of the segment polarity network in Drosophila. We explicitly describe the region in the parameter space where the wild-type expression pattern of key genes becomes feasible for each model. We also study how random walks in the parameter space exit from the feasible region of a network model, allowing us to compare the relative robustness of the alternative models.

@article{Dayarian2009Shape, abstract = {Author Summary Developing models with a large number of parameters for describing the dynamics of a biochemical network is a common exercise today. The dependence of predictions of such a network model on the choice of parameters is important to understand for two reasons. For the purpose of fitting biological data and making predictions, we need to know which combinations of parameters are strongly constrained by observations and also which combinations seriously affect a particular prediction. In addition, we expect naturally evolved networks to be somewhat robust to parameter changes. If the functioning of the network requires fine-tuning in many parameters, then mutations causing changes in regulatory interactions could quickly make the network dysfunctional. For predictions involving gene products being ON or OFF, we found a method that facilitates the study parameter dependence. As an example, we analyzed several competing models of the segment polarity network in Drosophila. We explicitly describe the region in the parameter space where the wild-type expression pattern of key genes becomes feasible for each model. We also study how random walks in the parameter space exit from the feasible region of a network model, allowing us to compare the relative robustness of the alternative models.}, added-at = {2019-08-05T10:16:49.000+0200}, author = {Dayarian, Adel and Chaves, Madalena and Sontag, Eduardo D. and Sengupta, Anirvan M.}, biburl = {https://www.bibsonomy.org/bibtex/2dfcb45643f8e725da42a15040ac85aa5/karthikraman}, description = {Shape, Size, and Robustness: Feasible Regions in the Parameter Space of Biochemical Networks}, doi = {10.1371/journal.pcbi.1000256}, interhash = {3892ed35cbdaa15c7d6f56841e72371a}, intrahash = {dfcb45643f8e725da42a15040ac85aa5}, journal = {PLOS Computational Biology}, keywords = {robustness}, month = {01}, number = 1, pages = {1-12}, publisher = {Public Library of Science}, timestamp = {2019-08-05T10:16:49.000+0200}, title = {Shape, Size, and Robustness: Feasible Regions in the Parameter Space of Biochemical Networks}, url = {https://doi.org/10.1371/journal.pcbi.1000256}, volume = 5, year = 2009 }