Predictive biophysical models of evolution

项目总结/摘要 拟议研究的总体目标是开发预测性多尺度生物物理 适应性进化动力学模型。在早期的工作中，我们证明了几种情况下， 生物医学的重要性，遗传变异的适应性效应可以准确地预测从一个 突变蛋白分子特性的独特组合。这一发现改变了 健身景观的概念从一个巧妙的比喻变成一个定量的易于处理的工具来预测 基因型-表型关系（GPR）。在此，我们以这些发现为基础， 进一步加深我们对生物物理和人口因素之间相互作用的理解， 决定适应性进化的动力和结果。我们将应用微流体技术， 自动化机器人设置沿着蛋白质工程和基因组编辑工具， 实验室实验中的进化动力学，在允许严格控制所有 规模-从分子到群体。为此，我们进行了一系列的进化 采用E.大肠杆菌逃避抗生素胁迫和结构 二氢叶酸还原酶的不稳定性。我们在所有尺度上刻画- 基因分型、分子性状、系统蛋白质组学和人口--多种进化途径， 新出现的细菌菌株的耐药性和适应性，并确定 描述（基因型，生物物理特性，系统响应）进化成为 可重复--也就是说可预测。与此同时， 使用多尺度模型，其中模型细胞的细胞质以生物病理学上真实的 模式生物的适应性是从其分子特征预测的， 实验性探地雷达突变体形式的分子性状使用最新技术预测 分子生物物理学的计算工具，允许复制和预测完整的 适应模型细胞群体的进化轨迹。紧密结合， 理论和实验将提供一个机会，发展预测的进化模型， 不断提高的准确性和真实性。沿着这些路线的进展将改变我们的方法 从描述性到预测性和定量的进化动力学研究， 有助于开发对抗抗生素耐药性的新方法， 潜在地，病毒逃避压力源，如药物和免疫反应。

Project Summary/Abstract The overarching goal of the proposed research is to develop predictive multiscale biophysical models of adaptive evolutionary dynamics. In earlier work we demonstrated for several cases of biomedical importance that fitness effect of genetic variation can be accurately predicted from a unique combination of molecular traits of the mutated protein. This finding transforms the concept of fitness landscape from an artful metaphor into a quantitative tractable tool to predict the genotype-phenotype relationship (GPR). Here we take these findings as a foundation to further extend our understanding of interplay between biophysical and population factors that determine the dynamics and outcome of adaptive evolution. We will apply microfluidics and automatic robotics setup along with protein engineering and genomic editing tools to explore evolutionary dynamics in laboratory experiments under conditions that allow tight control on all scales – from molecules to populations. To that end, we carry out a set of evolution experiments with adapting populations of E. coli escaping from antibiotic stress and structural instability of the essential protein Dihydrofolate Reductase. We characterize on all scales – genotyping, molecular traits, systems proteomics and population - multiple evolutionary paths to resistance and adaption of emerging bacterial strains and determine at which level of description (genotype, biophysical properties, systems responses) evolution becomes reproducible – and by implication predictable. In parallel we model the evolutionary dynamics using multiscale models where cytoplasm of model cells is presented in a biophysically realistic manner, and fitness of model organisms is predicted from its molecular traits using experimentally derived GPR. Molecular traits of mutant forms are predicted using state of the art computational tools of molecular biophysics allowing reproducing and predicting complete evolutionary trajectories of adapting populations of model cells. A tight integration between theory and experiment will provide an opportunity to develop predictive evolutionary models of ever increasing accuracy and realism. Progress along these lines will transform our approaches to study evolutionary dynamics from descriptive into predictive and quantitative, which will be instrumental to the development of novel approaches to fight antibiotic resistance and, potentially, viral escape from stressors such as drugs and immune response.