The oldest liverwort fossils are from the Late Devonian, ~360 MYA, the oldest mosses to be found in the fossil record are from the Permian, ~270 MYA .
The first deposits containing remnants of modern mosses are from the Jurassic and Cretaceous; based on these fossils some extant species exhibited only limited morphological change in the past 80 MY .
Angiosperms (flowering plants) are paleopolyploids, i.e. the genome of their common ancestor was subject to a large-scale or even genome-wide duplication event during the Late Jurassic or Early Cretaceous, 100–160 MYA .
Thus, “bryophytes” in comparison with vascular plants enable inference of early states of land plant evolution.Based upon spores found in the fossil record, the first plants had occupied the land in the Middle Ordovician, approximately 460 million years ago (MYA) [ (liverworts) from the remainder of the land plants, the vascular plants.All land plants display alternating multicellular generations – the sexual, haploid gametophyte and the asexual, diploid sporophyte.In early land plant fossils the gametophytic and sporophytic generation share about equal morphological complexity, making it likely that the gametophyte was reduced and the sporophyte became the dominant generation in vascular plants  while in “bryophytes” (mosses, hornworts and liverworts) the sporophyte generation was reduced and the gametophyte became dominant.The core eudicots apparently duplicated their genome in the Late Cretaceous, while the common ancestor of the ], prediction of open reading frames using species-specific models yielded a dataset of 22,237 coding sequences.
From those, a total of 2,907 paralogs were determined by all-against-all BLAST searches using previously described parameters [, poplar, or rice) as reference points.
Although using two different calibration dates (100 MYA for the -poplar split, and 150 MYA for the monocot-eudicot split) may affect the age distribution if one of the two calibration dates is unrealistic compared to the other, age distributions obtained for each calibration point separately were very similar (data not shown), suggesting that the dates of 100 MYA and 150 MYA  are in good agreement with inferred dates from tree topologies.
Sequences that were evolving too fast or too slow were removed, after which linearized trees, in which branch lengths are directly proportional to time, were constructed for each gene family [.
As can be clearly observed, a majority of the gene duplicates seem to have been created between 30–60 MYA (average 45 MYA), indicating that a large-scale gene duplication or a whole-genome duplication is indeed likely to have occurred around this time.
Based on the construction of linearized phylogenetic trees we infer the genome duplication to have occurred between 30 and 60 million years ago.
Gene Ontology and pathway association of the duplicated genes in ].