Limits and trade-offs of feedback control

 负反馈控制对于使生物系统对内部和外部扰动保持稳定至关重要，就像家庭需要恒温器来保持固定的室温一样。在细胞中，反馈被用来调节从基因表达到染色体复制的一切，它的失败会导致一系列人类疾病。但我们对生物学中反馈的理解是非常不完整的。大多数理论方法使用的数学框架不适合描述细胞，因为它们忽略了扰动的主要来源：化学反应涉及少量分子，并且由于每个反应都是概率性的，因此浓度波动会自发出现。与人类设计的系统不同，在人类设计的系统中，控制系统（例如恒温器）与它们控制的内容（温度波动）根本不同，细胞中的控制系统与它们控制的过程相似。这使得细胞很难抑制扰动，也使得研究人员很难分析控制系统的设计。在实验上，很少有系统允许精确测量单个细胞中的浓度和系统地修改控制系统以分析它们如何影响系统。分析也可能过于集中于一个特定系统的具体细节，以至于忽略或误解了一般指导原则。我们建议通过开发新的数学方法和系统地将我们的新实验测定应用于简单的模型系统，细菌质粒来解决这些问题。我们的初步理论证明了负反馈抑制细胞波动的能力的严格限制。它还提出了创造性和反直觉的机制，以尽量减少这些问题。值得注意的是，我们已经在质粒基因簇中发现了这些例子，我们知道这些基因簇受到强选择以抑制噪音。除了这些系统在理解反馈控制方面的有用性之外，细菌质粒在医学上也非常相关，因为它们导致了医院中的大多数耐药性病例，仅在美国，每年就有数百万人患上严重疾病，数万人死亡。它们也是生物技术行业的关键工具，我们研究的波动抑制特性是一个重大的麻烦。质粒无与伦比的实验易处理性使我们能够系统地改变控制特性，并通过实验严格测试数学描述，从而更深入地了解反馈机制，并增加有关质粒行为的有用知识。

 DESCRIPTION (provided by applicant): Negative feedback control is essential to make biological systems stable to internal and external perturbations, just as homes need thermostats to maintain a fixed room temperature. In cells feedback is used to regulate everything from gene expression to chromosome replication, and its failure causes a range of human disorders. But our understanding of feedback in biology is very incomplete. Most theoretical approaches use mathematical frameworks that are poorly suited to describe cells because they ignore a main source of perturbations: chemical reactions involve molecules in low numbers and since each individual reaction is probabilistic, fluctuations in concentrations arise spontaneously. Unlike human-designed systems, where the control systems (e.g. thermostats) are fundamentally different from what they control (temperature fluctuations), the control systems in cells are similar to the processes they control. This makes it very difficult for cells to suppress perturbations, and also makes it very difficult for researchers to analyze the control system design. Experimentally, very few systems allow both accurate measurements of concentrations in single cells and systematic modifications of the control system to analyze how they affect the system. Analyses can also focus so closely on the specific details of one specific system that general guiding principles are overlooked or misinterpreted. We propose to address these problems by developing new mathematical approaches and systematically applying our novel experimental assays to simple model systems, bacterial plasmids. Our preliminary theory demonstrates hard limits on the ability of negative feedback to suppress fluctuations in cells. It also suggests creative and counter-intuitive mechanisms that minimize these problems. Remarkably enough, we have found examples of these in plasmid gene clusters that we know are under strong selection to suppress noise. In addition to the usefulness of these systems to understand feedback control, bacterial plasmids are also very relevant medically, since they cause the majority of drug resistance cases in hospitals, a problem that leads to millions of serious illnesses and tens of thousands of deaths annually in the US alone. They are also key tools for the biotechnology industry, where the fluctuation suppression properties we study are a significant nuisance. The unmatched experimental tractability of plasmids allows us to systematically vary control properties and rigorously test the mathematical descriptions experimentally, leading to a deeper understanding of feedback mechanisms and also an increase in useful knowledge about plasmid behavior.