Fossil-calibrated Bayesian divergence dating across 195 plant taxa reveals the complementary power of concatenated rbcL+matK chloroplast markers
Abstract
This study examines species divergence time and phylogenetic relationships across 195 plant taxa using two widely used chloroplast molecular markers, rbcL and matK , analysed through Bayesian Inference in MrBayes under relaxed molecular clocks. Divergence time estimation was calibrated against seven fossil constraints, including Palaeonitella cranii (≥ 410 Ma), Rhyniophyta/ Cooksonia (420–430 Ma), and Archaefructus sinensis (125–130 Ma). Three datasets, rbcL , matK , and the concatenated rbcL + matK dataset, were analysed under the GTR + G + I substitution model to evaluate marker performance. The concatenated rbcL + matK dataset produced the deepest and most precisely constrained divergence estimates, dating land plant origin to 457.6 MYA, seed plant emergence to approximately 330 MYA, crown angiosperm radiation to 138.7 MYA, and the monocot–eudicot split to 125.0 MYA, all consistent with established fossil benchmarks. Sequence statistics revealed contrasting levels of nucleotide diversity between the two markers. Tajima’s D values differed among datasets, but because sampling was interspecific rather than population-based, we interpreted these statistics descriptively rather than as direct evidence of specific selective or demographic processes. The combined dataset displayed intermediate diversity (π = 0.1758) alongside narrower highest posterior density (HPD) intervals, reflecting improved estimation precision. The resulting time-calibrated phylogeny resolved 13 major clades spanning charophyte outgroups (447.60 MYA) through gymnosperms (412.25 MYA), pteridophytes, bryophytes, and monocots (395.08 MYA), to all eudicot superclades, providing a broad temporal framework across major geological intervals. Single-gene analyses revealed systematic limitations: rbcL underestimated deep divergence times, while matK generated wider HPD intervals at shallow nodes. The concatenated dataset generally showed stronger node support, fewer apparent polytomies, and narrower 95% HPD intervals than the individual-marker analyses, indicating improved phylogenetic resolution and greater statistical precision under the analytical framework examined here. These findings support the comparative use of rbcL and matK for broad plastid phylogenetic and divergence-time analyses, while highlighting the influence of marker choice, fossil calibration, taxon sampling, and model assumptions on estimated evolutionary chronologies, offering important implications for plant systematics, biogeography, macro-evolutionary studies, and biodiversity conservation.