Using Simulations to Explore Genetic Networks

正在开发的计算机软件采用基因网络的非数学描述，将其转换为一组微分方程，然后求解这些方程，以揭示网络成分的浓度如何随时间变化。该程序包括搜索这些参数值的工具，这些参数值使网络产生所需的时空模式。一条大大加快搜寻速度的捷径已经找到了。从一个简单的模型开始，只涉及果蝇的五个片段极性基因的产物，已经发现这些基因之间最具特征的相互作用不能解释片段极性模式。这个最初的失败指向了几个意想不到的实验结果，结果网络对参数值的变化具有惊人的鲁棒性。该网络还可以很容易地产生除生物片段极性模式外的几种特征模式，并具有不同的参数值。这两个结果都可能具有进化意义。这项初步工作将扩大到包括下列目标：A)找出迄今发现的转录控制和转录后修饰机制如何以及是否结合起来解释该网络在野生型和突变型苍蝇胚胎胚皮中观察到的时空模式行为，b)实验测量对该模型行为产生关键影响的某些参数的值，c)实验测试模型确定的某些转录调节步骤作为所有重要环节，d)理解如何，在后来的发育和其他基因的参与下，相同的网络建立了影像学椎间盘的前/后腔室边界。一种计算机模拟方法集成交叉调控基因之间的已知相互作用到定量机制模型已经开发。将进行进一步的改进和实验测试。将继续开发和及时提供将作为发育生物学家研究和教学工具的软件。

Computer software being developed takes a non-mathematical description of a gene network, converts it into a set of differential equations, and then solves the equations to reveal how concentrations of network components change over time. The program includes facilities to search for values of these parameters that cause the network to produce the desired spatio-temporal pattern. A short cut which greatly speeds up the search has been found. Starting with a simple model involving products of just five of the segment polarity genes of Drosophila, it has been found that the best-characterized interactions between those genes could not account for the segment polarity pattern. This initial failure pointed to several unexpected experimental results and the resulting network is surprisingly robust to changes in parameter values. The network can also readily produce, with different parameter values, a few characteristic patterns other than the biological segment polarity pattern. Both of these results may have evolutionary implications. This preliminary work will be extended to encompass the following objectives: a) to find out how and if the transcriptional control and post-transcriptional modification mechanisms so far discovered combine to explain the observed spatio-temporal patterning behavior of this network in wild type and mutant fly embryo blastoderms, b) to experimentally measure the values of certain parameters that affect crucially this model's behavior, c) to experimentally test certain transcription regulation steps pinpointed by the model as all important links, d) to comprehend how, later in development and modified by participation of other genes, the same network establishes the anterior/posterior compartment boundary in imaginal discs.A computer simulation approach for integrating known interactions among cross-regulatory genes into a quantitative mechanistic model has been developed. Further refinement and experimental testing will be done. Software that will serve as a tool for developmental biologists in both research and teaching will be continued to be developed and made available in a timely fashion.