Adaptation by improbable changes can identify the phenotypic targets of selection

Objective.我的实验室将开发，测试和应用统计方法来确定哪些表型变化是分子适应过程中选择的目标。理由。适应是进化的特征，因为选择改善了它们功能的一个方面，确定突变的功能效应是现代生物学研究的主要目标之一。然而，即使是详细的功能知识也不足以证明适应性。很难确定哪些进化变化是适应，因为突变经常，或者可能总是，影响一个以上的特征。因此，即使一个性状是通过改变某些测量功能的突变而进化的，选择的真实的目标可能是这些相同的突变对相关性状的影响。Approach.生物体的许多特征显然是不可能的。复杂的或精细优化的特征需要适应性解释，因为选择是唯一能够将系统一致地推向不可能状态的机制。基于这种理解，我的研究计划的长期目标是开发方法来识别适应，而不是通过它们的功能，而是通过它们发生的可能性。我提出了两个短期目标来支持这一长期目标。目标1.我们将通过预测生物物理蛋白质表型来测试鲸鱼肌红蛋白的适应性。潜水哺乳动物将氧气储存在肌肉肌红蛋白中。我们将测试两个相互竞争的适应性故事，肌红蛋白如何进化到耐受高肌肉浓度而不沉淀：电荷或稳定性的增加。这两种特征中的任何一种是否朝着一个不太可能的方向进化？这两种特性都可以从肌红蛋白的生物物理特性中预测出来。所以我们可以为一个等位基因的所有单个突变生成表型，并估计表型随机向一个方向或另一个方向变化的概率。这些完全解析的表型空间也提供了探索适应理论基础的令人兴奋的项目。目标二。我们将测试一个经典的例子发光颜色适应突变表型的大规模实证测量。点击甲虫的荧光素酶发光已经从绿色反复进化到橙子。这种变化是常见还是罕见？我们将通过表达人工突变体，根据等位基因的颜色将其分类到池中，并测序来评估荧光素酶的表型空间。深度测序的等位基因池（每个池具有不同的表型）也可以用于训练用于表型预测的机器学习算法。重要性适应是生物学的中心，但我们需要知道选择作用于哪些特征，以了解进化的原因。分子适应性的严格统计检验的应用遍及整个生物学领域。从追踪适应功能（作物或变化的环境）或适应不良功能（疾病的多效性原因）的起源，到了解冲突如何驱动新奇（免疫识别或物种形成）。

Objective. My lab will develop, test, and apply statistical methods to identify which phenotypic changes were targets of selection during molecular adaptation. Rationale. Adaptations are traits that evolved because selection improved an aspect of their function, and determining the functional effects of mutations is one primary goal of modern biological research. However, even detailed knowledge of function is not sufficient to demonstrate adaptation. It is difficult to determine which evolutionary changes are adaptations because mutations often, or perhaps always, influence more than one trait. So, even if a trait evolves by mutations that change some measured function, the real target of selection could be the effects of these same mutations on a correlated trait. Approach. Many features of organisms are manifestly improbable. Complex or finely-optimized traits invite adaptive explanations because selection is the only mechanism that can drive systems consistently toward unlikely states. Based on this understanding, my research program's long-term goal is to develop methods to identify adaptations not by their function but by the improbability of their occurrence. I propose two short-term aims to support this long-term goal. Aim 1. We will test proposed adaptations of whale myoglobin by predicting biophysical protein phenotypes. Diving mammals hold oxygen in muscle myoglobin. We will test two competing adaptive stories of how myoglobin evolved to tolerate high muscle concentrations without precipitating: increases in charge or stability. Did either trait evolve in an improbable direction? Both traits can be predicted from the biophysical properties of myoglobin. So we can generate phenotypes for all single mutations of an allele, and estimate the probability that a phenotype will change in one direction or the other by chance. These fully-resolved phenotype spaces also offer exciting projects that explore the theoretical basis of adaptation. Aim 2. We will test a classic example of luminescent colour adaptation by large-scale empirical measurement of mutant phenotypes. Click-beetle luciferase glow has evolved repeatedly from green to orange. Is this change common or rare? We will evaluate the phenotype space for luciferase by expressing artificial mutants, sorting alleles into pools based on their colour, and sequencing. Deep-sequenced pools of alleles, each pool a different phenotype, may also be useful for training machine-learning algorithms for phenotype prediction. Importance. Adaptation is the center of biology, but we need to know which traits selection acts on to understand the causes of evolution. The applications of a rigorous statistical test of molecular adaptation span all of biology. From tracing the origin of adaptive functions (in crops or changing environments) or maladaptive functions (in the pleiotropic causes of disease) to understanding how conflicts drive novelty (in immune recognition or species formation).