了解进化中突变和重组事件的频率和相对权重对于理解微生物如何达到拟合表型至关重要。传统上,这些进化参数是通过使用来自多位点序列分型(MLST)的数据推断的,众所周知,这产生了相互矛盾的结果。在不久的将来,这些估计肯定会通过全基因组序列的计算分析来进行.然而,不知道这种方法是否会产生准确的结果,因为细菌基因组表现出基因座类别的异质性表示,目前尚不清楚基因座性质如何影响此类估计。因此,我们评估了突变和重组推断是如何由具有不同遗传特征的基因座形成的,以沙眼衣原体为研究模型。我们发现,分配大量等位基因和正选择基因的基因座产生了非收敛估计和不一致的系统发育,因此更容易出现混淆算法。出乎意料的是,对于正在评估的模型,管家基因和非编码区以类似的方式形成了估计,这指出了后者在沙眼衣原体进化中的非随机作用。尽管目前的结果与一种特定的细菌有关,我们推测微生物特异性基因组结构(如编码能力,多态性分散,和正选择的基因座的分数)可以在估计重组和突变率时差异缓冲混杂因素的影响,因此,影响为此目的使用全基因组序列的准确性。在讨论通过全基因组序列分析获得的结果时,应考虑与计算机推理相关的这种推定偏差。其中“一刀切”的方法可能不适用。
The knowledge of the frequency and relative weight of mutation and recombination events in evolution is essential for understanding how microorganisms reach fitted phenotypes. Traditionally, these evolutionary parameters have been inferred by using data from multilocus sequence typing (MLST), which is known to have yielded conflicting results. In the near future, these estimations will certainly be performed by computational analyses of full-genome sequences. However, it is not known whether this approach will yield accurate results as bacterial genomes exhibit heterogeneous representation of loci categories, and it is not clear how loci nature impacts such estimations. Therefore, we assessed how mutation and recombination inferences are shaped by loci with different genetic features, using the bacterium Chlamydia trachomatis as the study model. We found that loci assigning a high number of alleles and positively selected genes yielded nonconvergent estimates and incongruent phylogenies and thus are more prone to confound algorithms. Unexpectedly, for the model under evaluation, housekeeping genes and noncoding regions shaped estimations in a similar manner, which points to a nonrandom role of the latter in C. trachomatis evolution. Although the present results relate to a specific bacterium, we speculate that microbe-specific genomic architectures (such as coding capacity, polymorphism dispersion, and fraction of positively selected loci) may differentially buffer the effect of the confounding factors when estimating recombination and mutation rates and, thus, influence the accuracy of using full-genome sequences for such purpose. This putative bias associated with in silico inferences should be taken into account when discussing the results obtained by the analyses of full-genome sequences, in which the \"one size fits all\" approach may not be applicable.