用简单微生物系统所做实验有可能为演化和生态过程提供线索。但这样的实验结果一般是否适用于真正的生态系统？Forde等人构建了关于宿主-寄生体共演化在不同营养水平上可能会怎样影响多样性的数学模型。利用T7噬菌体和大肠杆菌，他们发现，很多模拟结果对相互作用的生态细节是不敏感的。但该模型体系的特性是可以预测和通过实验确认的，说明这些数学模型是可靠的。这种将一系列模型与实验确认相结合的做法，为了解哪些预测也许会是模型敏感的、哪些更具一般意义提供了一个方法。（生物谷Bioon.com）
生物谷推荐原始出处：
Nature 455, 220-223 (11 September 2008) | doi:10.1038/nature07152
Understanding the limits to generalizability of experimental evolutionary models
Samantha E. Forde1,5, Robert E. Beardmore2,5, Ivana Gudelj2,3,5, Sinan S. Arkin2, John N. Thompson1 & Laurence D. Hurst4
1 Department of Ecology and Evolutionary Biology, University of California, Santa Cruz, California 95064, USA
2 Department of Mathematics, Imperial College London, London SW7 2AZ, UK
3 Department of Mathematical Sciences and,
4 Department of Biology & Biochemistry, University of Bath, Bath BA2 7AY, UK
5 These authors contributed equally to this work.
Given the difficulty of testing evolutionary and ecological theory in situ, in vitro model systems are attractive alternatives1; however, can we appraise whether an experimental result is particular to the in vitro model, and, if so, characterize the systems likely to behave differently and understand why? Here we examine these issues using the relationship between phenotypic diversity and resource input in the T7–Escherichia colico-evolving system as a case history. We establish a mathematical model of this interaction, framed as one instance of a super-class of host–parasite co-evolutionary models, and show that it captures experimental results. By tuning this model, we then ask how diversity as a function of resource input could behave for alternative co-evolving partners (for example, E. coli with lambda bacteriophages). In contrast to populations lacking bacteriophages, variation in diversity with differences in resources is always found for co-evolving populations, supporting the geographic mosaic theory of co-evolution2. The form of this variation is not, however, universal. Details of infectivity are pivotal: in T7–E. coli with a modified gene-for-gene interaction, diversity is low at high resource input, whereas, for matching-allele interactions, maximal diversity is found at high resource input. A combination of in vitro systems and appropriately configured mathematical models is an effective means to isolate results particular to the in vitro system, to characterize systems likely to behave differently and to understand the biology underpinning those alternatives.
