Genome size varies by orders of magnitude across land plants, and the factors driving evolutionary increases and decreases in genome size vary across lineages. Bryophytes have the smallest genomes relative to other land plants and there is growing evidence for frequent whole genome duplication (WGD) across the lineage. However, the broad scale patterns of genome size, chromosome number, and WGD have yet to be characterized across bryophytes in a phylogenetic context. In the present study, we use a phylogenetic comparative approach and leverage all published data on genome size, chromosome number, and WGD to reconstruct the evolutionary history of these traits across the three major bryophyte lineages: hornworts, liverworts, and mosses.  Each lineage of bryophytes exhibits a distinct pattern of genome size evolution and prevalence of WGD, with mosses having the most dynamic genome sizes and highest propensity for WGD. For the lineages with robust sampling, we found that 21.3% of mosses and 13% of liverworts species have naturally occurring polyploids. In addition, haploid genome size (1C) is most dynamic in the mosses which includes 15 transitions to larger genomes and nine reversals largely in the orders Dicranles and Hypnales. Crucially, there is no correlation between genome size and WGD or chromosome number, potentially suggesting rapid genome downsizing following WGD. As bryophytes are poikilohydric, desiccation tolerant plants, having large genomes may be physiologically prohibitive given the cost to growth and metabolism associated with them. These findings emphasize the unique evolution of the bryophytes broadly and of the hornworts, liverworts, and mosses individually, and should therefore serve as impetus for more in-depth experimental study of genome size evolution and WGD in bryophytes. 

Phylogenetic data collection - Pyphlawd

A DNA matrix of 3,890 taxa and 10 concatenated loci (5.8S, 26S, atpb, nad5, psbA-trnH, psbA, rbcL, rps4-trnS, trnA, trnL-F) was assembled using PyPhlawd (Smith & Walker, 2018). PyPhlawd was run using “Embryophyte” as the taxonomic group and then narrowed to all  bryophyte species (hornworts, liverworts, and mosses) using a taxon list including all listed species available on genbank. The clustering function in PyPhlawd produced 311 clusters of which the ten with the largest taxonomic representation were selected for further analysis. 

 Phylogenetic analysis

Matrices for each locus were aligned using Mafft v7.419 with default settings. The loci were concatenated and each locus was assigned a separate partition. Phylogenetic inference was implemented in IQTREE v.2 (Minh et al., 2020) using Model Finder Plus (Kalyaanamoorthy et al., 2017) to find best fit models for each partition. IQTREE was implemented using 1000 bootstrap replicates. The concatenated matrix and phylogenetic tree are available on Dryad digital repository ()