Kuznetsov Gleb, Goodman Daniel B, Filsinger Gabriel T, Landon Matthieu, Rohland Nadin, Aach John, Lajoie Marc J, Church George M
Department of Genetics, Harvard Medical School, Boston, MA, USA.
Wyss Institute for Biologically Inspired Engineering, Harvard Medical School, Boston, MA, USA.
Genome Biol. 2017 May 25;18(1):100. doi: 10.1186/s13059-017-1217-z.
We present a method for identifying genomic modifications that optimize a complex phenotype through multiplex genome engineering and predictive modeling. We apply our method to identify six single nucleotide mutations that recover 59% of the fitness defect exhibited by the 63-codon E. coli strain C321.∆A. By introducing targeted combinations of changes in multiplex we generate rich genotypic and phenotypic diversity and characterize clones using whole-genome sequencing and doubling time measurements. Regularized multivariate linear regression accurately quantifies individual allelic effects and overcomes bias from hitchhiking mutations and context-dependence of genome editing efficiency that would confound other strategies.
我们提出了一种通过多重基因组工程和预测建模来识别优化复杂表型的基因组修饰的方法。我们应用该方法识别出六个单核苷酸突变，这些突变可恢复63密码子大肠杆菌菌株C321.∆A所表现出的59%的适应性缺陷。通过多重引入靶向变化组合，我们产生了丰富的基因型和表型多样性，并使用全基因组测序和倍增时间测量来表征克隆。正则化多元线性回归准确地量化了个体等位基因效应，并克服了来自搭便车突变的偏差以及基因组编辑效率的上下文依赖性，而这些会使其他策略混淆。
