Beckmann John, Gillespie Joe, Tauritz Daniel
Auburn University Department of Entomology and Plant Pathology, 301 Funchess Hall, Auburn, AL; 36849.
University of Maryland Baltimore, School of Medicine, Department of Microbiology and Immunology, Baltimore, 685 W. Baltimore St., HSF I Suite 380, Baltimore, MD 21201.
bioRxiv. 2023 Mar 25:2023.03.23.533954. doi: 10.1101/2023.03.23.533954.
Evolutionary algorithms (EAs) simulate Darwinian evolution and adeptly mimic natural evolution. Most EA applications in biology encode high levels of abstraction in top-down ecological population models. In contrast, our research merges protein alignment algorithms from bioinformatics into codon based EAs that simulate molecular protein string evolution from the bottom up. We apply our EA to reconcile a problem in the field of induced cytoplasmic incompatibility (CI). is a microbial endosymbiont that lives inside insect cells. CI is conditional insect sterility that operates as a toxin antidote (TA) system. Although, CI exhibits complex phenotypes not fully explained under a single discrete model. We instantiate in-silico genes that control CI, CI factors ( ), as strings within the EA chromosome. We monitor the evolution of their enzymatic activity, binding, and cellular localization by applying selective pressure on their primary amino acid strings. Our model helps rationalize why two distinct mechanisms of CI induction might coexist in nature. We find that nuclear localization signals (NLS) and Type IV secretion system signals (T4SS) are of low complexity and evolve fast, whereas binding interactions have intermediate complexity, and enzymatic activity is the most complex. Our model predicts that as ancestral TA systems evolve into eukaryotic CI systems, the placement of NLS or T4SS signals can stochastically vary, imparting effects that might impact CI induction mechanics. Our model highlights how preconditions, genetic diversity, and sequence length can bias evolution of towards one mechanism or another.
进化算法（EAs）模拟达尔文进化过程，并巧妙地模仿自然进化。生物学中大多数进化算法应用都在自上而下的生态种群模型中编码了高度抽象的信息。相比之下，我们的研究将生物信息学中的蛋白质比对算法融入基于密码子的进化算法中，从下至上模拟分子蛋白质序列的进化。我们应用我们的进化算法来解决诱导细胞质不相容性（CI）领域中的一个问题。沃尔巴克氏体是一种生活在昆虫细胞内的微生物内共生体。细胞质不相容性是一种有条件的昆虫不育现象，作为一种毒素解毒剂（TA）系统发挥作用。尽管如此，细胞质不相容性表现出的复杂表型在单一离散模型下并未得到充分解释。我们在计算机模拟中，将控制细胞质不相容性的基因，即细胞质不相容性因子（Cif），实例化为进化算法染色体中的序列。我们通过对其一级氨基酸序列施加选择压力，监测它们的酶活性、结合能力和细胞定位的进化情况。我们的模型有助于解释为什么两种不同的细胞质不相容性诱导机制可能在自然界中共存。我们发现核定位信号（NLS）和IV型分泌系统信号（T4SS）的复杂性较低且进化迅速，而结合相互作用具有中等复杂性，酶活性则最为复杂。我们的模型预测，随着祖先TA系统进化为真核细胞质不相容性系统，核定位信号或IV型分泌系统信号的位置可能会随机变化，从而产生可能影响细胞质不相容性诱导机制的效应。我们的模型强调了前提条件、遗传多样性和序列长度如何使沃尔巴克氏体的进化偏向一种或另一种机制。
