Pereira Rui, Vilaça Paulo, Maia Paulo, Nielsen Jens, Rocha Isabel
CEB - Centre of Biological Engineering , University of Minho, Campus de Gualtar , Braga 4710-057 , Portugal.
Department of Biology and Biological Engineering , Chalmers University of Technology , SE412 96 Gothenburg , Sweden.
ACS Synth Biol. 2019 May 17;8(5):976-988. doi: 10.1021/acssynbio.8b00248. Epub 2019 Apr 15.
The uncertain relationship between genotype and phenotype can make strain engineering an arduous trial and error process. To identify promising gene targets faster, constraint-based modeling methodologies are often used, although they remain limited in their predictive power. Even though the search for gene knockouts is fairly established in constraint-based modeling, most strain design methods still model gene up/down-regulations by forcing the corresponding flux values to fixed levels without taking in consideration the availability of resources. Here, we present a constraint-based algorithm, the turnover dependent phenotypic simulation (TDPS) that quantitatively simulates phenotypes in a resource conscious manner. Unlike other available algorithms, TDPS does not force flux values and considers resource availability, using metabolite production turnovers as an indicator of metabolite abundance. TDPS can simulate up-regulation of metabolic reactions as well as the introduction of heterologous genes, alongside gene deletion and down-regulation scenarios. TDPS simulations were validated using engineered Saccharomyces cerevisiae strains available in the literature by comparing the simulated and experimental production yields of the target metabolite. For many of the strains evaluated, the experimental production yields were within the simulated intervals and the relative strain performance could be predicted with TDPS. However, the algorithm failed to predict some of the production changes observed experimentally, suggesting that further improvements are necessary. The results also showed that TDPS may be helpful in finding metabolic bottlenecks, but further experiments would be required to confirm these findings.
基因型与表型之间不确定的关系可能使菌株工程成为一个艰巨的反复试验过程。为了更快地识别有前景的基因靶点，人们经常使用基于约束的建模方法，尽管其预测能力仍然有限。尽管在基于约束的建模中寻找基因敲除已经相当成熟，但大多数菌株设计方法在对基因上调/下调进行建模时，仍通过将相应的通量值强制设定为固定水平来实现，而没有考虑资源的可用性。在此，我们提出一种基于约束的算法——周转率依赖性表型模拟（TDPS），该算法以一种资源意识的方式定量模拟表型。与其他现有算法不同，TDPS不强制通量值，并考虑资源可用性，使用代谢物生产周转率作为代谢物丰度的指标。TDPS可以模拟代谢反应的上调以及异源基因的引入，同时还能模拟基因删除和下调的情况。通过比较目标代谢物的模拟产量和实验产量，使用文献中现有的工程酿酒酵母菌株对TDPS模拟进行了验证。对于许多评估的菌株，实验产量在模拟区间内，并且可以用TDPS预测相对菌株性能。然而，该算法未能预测一些实验观察到的产量变化，这表明有必要进一步改进。结果还表明，TDPS可能有助于找到代谢瓶颈，但需要进一步的实验来证实这些发现。
