Data from: Evaluating the impact of genomic data and priors on Bayesian estimates of the angiosperm evolutionary timescale

The evolutionary timescale of angiosperms has long been a key question in biology. Molecular estimates of this timescale have shown considerable variation, being influenced by differences in taxon sampling, gene sampling, fossil calibrations, evolutionary models, and choices of priors. Here, we analyse a data set comprising 76 protein-coding genes from the chloroplast genomes of 195 taxa spanning 86 families, including novel genome sequences for 11 taxa, to evaluate the impact of models, priors, and gene sampling on Bayesian estimates of the angiosperm evolutionary timescale. Using a Bayesian relaxed molecular-clock method, with a core set of 35 minimum and two maximum fossil constraints, we estimated that crown angiosperms arose 221 (251–192) million years ago during the Triassic. Based on a range of additional sensitivity and subsampling analyses, we found that our date estimates were generally robust to large changes in the parameters of the birth-death tree prior and of the model of rate variation across branches. We found an exception to this when we implemented fossil calibrations in the form of highly informative gamma priors rather than as uniform priors on node ages. Under all other calibration schemes, including trials of seven maximum age constraints, we consistently found that the earliest divergences of angiosperm clades substantially predate the oldest fossils that can be assigned unequivocally to their crown group. Overall, our results and experiments with genome-scale data suggest that reliable estimates of the angiosperm crown age will require increased taxon sampling, significant methodological changes, and new information from the fossil record.被子植物的演化时间尺度长期以来都是生物学领域的核心科学议题。针对该时间尺度的分子估计结果存在显著差异，其差异受到类群取样、基因取样、化石校准、演化模型以及先验选择等多方面因素的影响。本研究分析了一套数据集，包含覆盖86个科的195个类群的叶绿体基因组（chloroplast genomes）中的76个蛋白质编码基因（protein-coding genes），其中包含11个类群的全新基因组序列，以此评估模型、先验以及基因取样对被子植物演化时间尺度贝叶斯估计的影响。研究采用贝叶斯松弛分子钟方法（Bayesian relaxed molecular-clock method），结合包含35个最小校准点与2个最大校准点的核心化石约束，估计得出被子植物冠群起源于三叠纪，距今2.21亿年（95%置信区间：1.92亿至2.51亿年）。基于一系列额外的敏感性分析与子取样分析，本研究发现该年代估计结果总体上对出生-死亡树先验（birth-death tree prior）的参数以及分支间速率变异模型的大幅调整具有稳健性。但存在一个例外：当将化石校准设置为高信息量伽马先验（gamma priors）而非节点年龄的均匀先验（uniform priors）时，估计结果出现偏差。在所有其他校准方案中，包括7个最大年龄约束的测试，我们一致发现被子植物支系的最早分化时间显著早于可明确归为冠群的最古老化石记录。总体而言，本研究结果以及基因组规模数据的相关实验表明，要获得可靠的被子植物冠群年龄估计，需要扩大类群取样范围、开展重大的方法学改进，并获取化石记录的新信息。