作者: Mark P. Simmons , John Gatesy
DOI: 10.1016/J.YMPEV.2015.05.011
关键词:
摘要: It has recently been concluded that phylogenomic data from 310 nuclear genes support the clade of (Amborellales, Nymphaeales) as sister to remaining angiosperms and shortcut coalescent phylogenetic methods outperformed concatenation for these data. We falsify both those conclusions here by demonstrating discrepant results between analyses are primarily caused applied (MP-EST STAR) not being robust highly divergent often mis-rooted gene trees were used. This result reinforces expectation low amounts signal methodological artifacts in gene-tree reconstruction can be more problematic than is assumption a single hierarchy all when approaches ancient divergences empirical studies. also demonstrate third method, ASTRAL, MP-EST or STAR, Observed Variability (OV) Tree Independent Generation Evolutionary Rates (TIGER), which two character subsampling procedures, biased favor characters with asymmetrical distributions states this dataset. conclude enthusiastic application novel tools substitute rigorous first principles, trending (e.g., divergences, tree-independent subsampling), may sources previously under-appreciated, systematic errors.