Is (non)parallel evolution driven by (non)parallel selection?

进化有多可预测？在大致相似的环境条件下，自然选择是否总是倾向于相似的性状和基因？如果进化是可预测的，那么人类将能够更好地预测疾病、农业害虫或我们希望在面对环境变化时保护的物种的进化变化。这项研究将检验进化可预测的程度。具体来说，研究人员将测试在许多复制种群中某些特征的重复进化是否证明了类似的自然选择，正如人们普遍认为的那样，但从未在自然界中得到验证。这项工作范围广泛，整合了环境、生态、形态学、免疫学和基因组数据，因此将为未来的进化研究提供基准。此外，研究的生物——三刺鱼——数量丰富，种类繁多，研究得很充分，这使它成为向K-12科学教师和学生教授生物学的绝佳工具。该研究提出了以下问题：(1)在我们观察到重复进化的地方，自然选择是否也非常相似？(2)免疫特性是否以可预测的方式进化？(3)平行自然选择是否作用于相同的基因组区域？(4)我们如何彻底地解释遗传对选择的反应？为了回答这些问题，研究人员将把湖水鱼和溪流鱼之间的F2杂交种放入自然栖息地的围栏中，以测量许多性状和基因组位点的选择。通过在四个独立的湖泊-溪流对中重复这些选择措施，他们可以首次评估野外平行选择与平行进化之间的关系。

How predictable is evolution? Given broadly similar environmental conditions, does natural selection always favor similar traits and genes? If evolution is predictable, then humanity will be better able to anticipate evolutionary changes in diseases, agricultural pests, or species we wish to preserve in the face of environmental change. This study will examine the degree to which evolution is predictable. Specifically, the researchers will test whether the repeated evolution of certain traits in many replicate populations is evidence of similar natural selection, as has been widely assumed but never tested in nature. The work has broad scope, integrating environmental, ecological, morphological, immunological, and genomic data, and so will provide a benchmark for future studies of evolution. Moreover, the study organism -- the threespine stickleback fish -- is abundant, diverse, and well-studied, making it an excellent tool for teaching biology to both K-12 science teachers and their students.The study asks the following questions: (1) where we observe repeated evolution, is natural selection also very similar?, (2) do immune traits evolve in a predictable manner?, (3) does parallel natural selection act on the same genomic regions?, (4) how thoroughly can we explain genetic responses to selection? To answer these questions, the researchers will place F2 hybrids between lake and stream fish into enclosures in natural habitats, to measure selection on many traits and genomic loci. By repeating these measures of selection in four independent lake-stream pairs, they can, for the first time, evaluate the relationship between parallel selection and parallel evolution in the wild.