Department of Biomedical Engineering, University of Virginia, Charlottesville, VA 22908, USA.
Bioinformatics. 2011 Oct 15;27(20):2888-94. doi: 10.1093/bioinformatics/btr496. Epub 2011 Aug 31.
Biological networks are robust to a wide variety of internal and external perturbations, yet fragile or sensitive to a small minority of perturbations. Due to this rare sensitivity of networks to certain perturbations, it is unclear how precisely biochemical parameters must be experimentally measured in order to accurately predict network function.
Here, we examined a model of cardiac β-adrenergic signaling and found that its robustness portrait, a global measure of steady-state network function, was well conserved even when all parameters were rounded to their nearest 1-2 orders of magnitude. In contrast, β-adrenergic network kinetics were more sensitive to parameter precision. This analysis was then extended to 10 additional networks, including Escherichia coli chemotaxis, stem cell differentiation and cytokine signaling, of which nine exhibited conserved robustness portraits despite the order-of-magnitude approximation of their biochemical parameters. Thus, both fragile and robust aspects of diverse biological networks are largely shaped by network topology and can be predicted despite order-of-magnitude uncertainty in biochemical parameters. These findings suggest an iterative strategy where order-of-magnitude models are used to prioritize experiments toward the fragile network elements that require precise measurements, efficiently driving model revision.
Supplementary data are available at Bioinformatics online.
生物网络对各种内部和外部的扰动具有很强的鲁棒性，但对少数扰动却很脆弱或敏感。由于网络对某些扰动的这种罕见敏感性，目前尚不清楚为了准确预测网络功能，必须在多大程度上精确测量生化参数。
在这里，我们研究了心脏β-肾上腺素能信号转导的模型，发现即使将所有参数四舍五入到最接近的 1-2 个数量级，其稳态网络功能的全局度量，即稳健性特征仍然很好地保持。相比之下，β-肾上腺素能网络动力学对参数精度更敏感。然后将该分析扩展到 10 个额外的网络，包括大肠杆菌趋化性、干细胞分化和细胞因子信号转导，其中 9 个网络尽管其生化参数的数量级近似，但仍具有保守的稳健性特征。因此，不同生物网络的脆弱和稳健方面在很大程度上都由网络拓扑结构决定，尽管生化参数存在数量级的不确定性，但仍可以进行预测。这些发现表明，可以采用迭代策略，使用数量级模型来优先考虑需要精确测量的脆弱网络元素的实验，从而有效地推动模型修正。
补充数据可在“生物信息学”在线获取。
