Tripods+X:Res:Collaborative Research: Identification of Gene Regulatory Network Function from Data

分子和细胞生物学处于我们理解生命和改善人类健康的前沿。然而，关于基因和信号如何实时相互作用以产生细胞行为的细节还没有被很好地理解。问题是，缺乏将实验数据整合到描述系统重要行为的模型中，但又不描述不必要的细节的方法，这将使它在合理的时间内计算过于繁琐。这个项目将开发一套新的工具，可以预测相互依赖的基因网络的整个行为范围，并以如此有效的方式进行，以便可以很容易地探索几个突变或可选基因集行为的影响下的行为。此外，一旦一组基因的行为有了这样的描述，它就会被其他科学家分享和使用。这些结果的不断积累将迅速增加它们对数据科学、细胞生物学和更广泛科学的价值。自从科学获得了对基因组进行测序的能力以来，分子和细胞生物学已经见证了巨大的飞跃。然而，虽然基因组是相对静态的，但表型是许多基因相互作用的结果，这些基因在时间上动态表达。此外，由于表型来自复杂的转录网络，其中的相互作用往往是非线性的，基于数据的模型将在理解和最终控制这些表型方面发挥关键作用。目前的建模技术受到物理学的推动，努力弥合告知模型参数值的生物测量的低分辨率与动力系统对初始条件和参数敏感的基本事实之间的根本冲突。这个项目开发了一个新的数学框架，它提供了全球动态的定量描述，与粗糙的、有噪音的生物测量相兼容。与用于构建数据库的计算高效算法相结合，该数据库可以捕获网络中与生物相关的动态，它可以成为基因网络空间中共享科学的基础。这些工具将用于根据实验数据构建和验证网络模型，询问给定网络的动态行为，并比较网络的动态摘要。该奖项反映了NSF的法定使命，并通过使用基金会的智力优势和更广泛的影响审查标准进行评估，被认为值得支持。

Molecular and cellular biology are at the forefront of our quest to understand life and improve human health. However, the details of how genes and signals interact in real time to produce cell behavior, are not well understood. The problem is a lack of approaches that can integrate experimental data into models that describe the important behavior of the system, yet do not describe nonessential details, that would make it too cumbersome to compute in a reasonable time. This project will develop a new set of tools, that can predict the entire range of behaviors of a network of interdependent genes and do it so efficiently that behavior under the effect of a few mutations or behavior of alternative sets of genes can be readily explored. Furthermore, once such a description of behavior of a set of genes is available, it will be shared and used by other scientists. Continuous build-up of these results will rapidly increase their value for data science, cellular biology, and broader science. Molecular and cellular biology have witnessed a huge leap forward since science has acquired the ability to sequence genomes. However, while genomes are relatively static, the phenotypes result from the interaction of many genes that are expressed dynamically in time. Furthermore, since phenotypes arise from complex transcriptional networks, where the interactions tend to be nonlinear, data-based models will play a key role in understanding and ultimately controlling these phenotypes. Current modeling techniques, motivated by physics, struggle to bridge a fundamental conflict between low resolution of biological measurements informing model parameter values and the fundamental fact that dynamical systems are sensitive to initial conditions and parameters. This project develops a novel mathematical framework that provides a quantitative description of global dynamics that are compatible with coarse, noisy biological measurements. Combined with computationally efficient algorithms for the construction of databases that capture biologically relevant dynamics of a network, it can become the basis for shared science in the space of gene networks. These tools will be used to construct and validate network models from experimental data, interrogate dynamic behavior of a given network, and compare dynamics summaries across networks.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.