Jamshidi Neema, Vo Thuy D, Palsson Bernhard O
Department of Bioengineering, University of California, San Diego, CA, USA.
Methods Mol Biol. 2007;366:267-85. doi: 10.1007/978-1-59745-030-0_15.
The availability and accessibility of high-throughput and biological legacy data have allowed mathematical analyses of genome-scale metabolic networks and models. Model formulation is centered on the conservation principles of mass and charge. Thermodynamic information is generally incorporated by means of reaction reversibility. If further experimental data are available, such as kinetic parameters, models describing system evolution over time can be developed. The type of data available largely determines the type of model (and subsequently the type of analysis) that can be performed. Different modeling approaches offer different advantages. Detailed kinetic models can make specific predictions about network functional states given knowledge about the enzyme parameter variations resulting from single-nucleotide polymorphisms (SNPs). They also require a large amount of experimental data, which is rarely available. On the other hand, although current formulations using the constraint-based optimization framework do not offer information about metabolite concentrations or time-dependent changes, it is a remarkably flexible modeling framework and permits the integration of a large amount of very different data types.
高通量和生物学遗留数据的可得性与可获取性，使得对基因组规模代谢网络和模型进行数学分析成为可能。模型构建以质量和电荷守恒原理为核心。热力学信息通常通过反应可逆性来纳入。如果有更多实验数据可用，比如动力学参数，那么就可以开发描述系统随时间演变的模型。可用数据的类型在很大程度上决定了能够执行的模型类型（以及随后的分析类型）。不同的建模方法具有不同的优势。给定关于单核苷酸多态性（SNP）导致的酶参数变化的知识，详细的动力学模型能够对网络功能状态做出具体预测。它们也需要大量实验数据，而这些数据很少能获得。另一方面，尽管当前基于约束的优化框架的公式化方法无法提供关于代谢物浓度或随时间变化的信息，但它是一个非常灵活的建模框架，允许整合大量非常不同的数据类型。
