Probabilistic approaches to optimize synthetic organisms

合成生物学是基于生物系统的“正向工程”：从定义明确的遗传成分的工具箱中，基本的构建模块被用来创建自然界中不存在的组合。合成生物学越来越多地用于复杂分子的合成要么过于昂贵，要么不可能在实验室或工厂中创建。设计新的或重新设计的酶途径进入微生物有望成为一种具有成本效益的替代方案。然而，这仍然是具有挑战性的，即使在遗传上易于处理的生物体，如面包酵母酿酒酵母。由此产生的系统通常是功能性的，但远非最佳。 系统生物学基于生物系统的“逆向工程”。在这里，高通量组学技术（例如下一代测序）通常用于分析大量靶生物样品。通常，目标是从这些特征中推断出潜在的生物机制。高通量分析和生物信息学分析的共同使用通过详细描述单个基因、途径和过程的状态，提供了对生物体状态的更全面的分子视角。 该项目将建立系统生物学模型，以帮助创造合成生物体。概率模型用于在基因型（诱导性状的遗传扰动）和表型（目标性状的表达）之间建立“中介”。所谓的转录签名网络（TSN）允许我们在迭代干扰合成生物体以提高功效时，对合成生物体的全局状态进行推理。在每个设计步骤中考虑系统分子状态的能力将使我们能够在正向和反向工程之间无缝切换。 我们通过优化现有的合成酵母系统来证明我们的方法的概念，该系统产生苄基异喹啉生物碱，这是一类植物次生代谢产物，包括止咳，抗菌和抗肿瘤药物。当前系统的产量对于实际应用来说太低，这是合成生物学中的常见情况。一个半理性的定向进化方法与我们的TSN系统地扰动合成系统，提高生物碱产量。 这一高度跨学科的研究成果将对加拿大产生许多好处。我们的短期目标是优化现有的生物燃料和药物微生物工厂，这可能会产生积极的商业影响。此外，这些独特的，复杂的化合物是研究的兴趣（例如替代阿片类药物的疼痛）。长期的智力挑战是建立预测模型，使微生物工厂的建设成为常规和自动化。这可能会带来全新的研究和工业机会。

Synthetic biology is based on the “forward engineering” of biological systems: from a toolbox of well-defined genetic components, the basic building blocks are used to create combinations not present in nature. Synthetic biology is increasingly used when the synthesis of a complex molecule is either prohibitively expensive or impossible to create in the laboratory or factory. The design of novel or re-engineered enzymatic pathways into microbes promises a cost effective alternative. However this remains challenging even within genetically tractable organisms such as baker's yeast Saccharomyces cerevisiae. The resultant systems are often functional but far from optimal. Systems Biology is based on the “reverse engineering” of biological systems. Here high-throughput -omic technologies (eg next generation sequencing) are typically used to profile a large number of samples of a target organism. Often the goal is to infer from these profiles the underlying biological machinery. The use of high-throughput profiling and bioinformatic analysis together provide a more holistic molecular perspective of the state of an organism by detailing the state of individual genes, pathways and processes. This project will build systems biology models to aid in the creation of synthetic organisms. Probabilistic models are used to build an “intermediary” between genotype (genetic perturbations to induce a trait) and phenotype (expression of the target trait). The so-called transcriptional signature network (TSN) allows us to reason about the global state of the synthetic organism as we iteratively perturb it to increase efficacy. The ability to consider the molecular state of the system at each design step will allow us to shift seamlessly between forward and reverse engineering. We show proof of concept for our approach by optimizing an existing synthetic yeast system that produces benzylisoquinoline alkaloids, a class of plant secondary metabolites that includes antitussive, antibacterial, and antineoplastic drugs. Yields of the current system are too low for practical applications, a common situation in synthetic biology. A semi-rational directed evolutionary approach is used with our TSN to systematically perturb the synthetic system and improve alkaloid yields. The outcomes of this highly interdisciplinary research will have many benefits to Canada. Our short term goals to optimize existing microbial factories for biofuels and drugs could have positive commercial impact. Moreover, these unique, complicated compounds are of interest in research (eg alternative opioids for pain). The longer term intellectual challenge is to build predictive models that will make microbial factory construction routine and automated. This could open entire new research and industrial opportunities.