Morphological evolution in mosses has long been hypothesized to accompany shifts in microhabitats and can be tested using comparative phylogenetics. These lines of inquiry have developed substantially, in part, by target capture sequencing allowing for phylogenomic scale data generated from herbarium specimens. In the present study, we test the relationship between taxonomically important morphological characters in the moss genus Fissidens, using both a 400-locus dataset generated using a target-capture approach as well as a three-locus phylogeny generated using sanger sequencing. Phylogenetic trees were generated using ASTRAL and Bayesian Inference and used to test the monophyly of subgenera/sections and provided the basis for ancestral character reconstruction and phylogenetic correlation analyses among five morphological characters as well as habitat moisture scored from literature. The characters axillary hyaline nodules, limbidium, costa, and peristome morphology as well as sexual system, minimum habitat moisture, average habitat moisture, maximum habitat moisture, and habitat moisture niche breadth each exhibit statistically significant phylogenetic signal. Significant correlations were found between the limbidium (phyllid/leaf border) and habitat moisture niche breadth, which could be interpreted as a more extensive limbidium enabling species to survive across a wider variety of habitats. Correlations were also found between costa anatomy and the limbidum of the gametophyte and sporophyte peristome morphology, as well as average habitat moisture and sexual system. Continued exploration of the relationships between morphological evolution, life history, and habitat will enable us to expand our understanding of functional morphology in mosses.

The three-gene dataset was generated by Sanger sequencing. Each locus was analyzed individually using maximum likelihood implemented in RAxML. These three genes were concatenated and analyzed using Bayesian inference implemented in Mr.Bayes. 

The ASTRAL dataset was generated by target capture methods using the GoFlag 451 and 408 loci probe sets. 404 loci were analyzed. We conducted phylogenetic analyses using two sets of data: (1) the probe dataset that includes only the conserved exon regions targeted by the sequencing probes, which are each, on average, 200 bp long and (2) the full dataset that included both the conserved exon regions from the probe dataset as well as the flanking intron regions. For both the probe and full dataset we conducted analyses of the concatenated supermatrix of all the loci using maximum likelihood implemented in RAxML. Additionally, each locus was analyzed individually using maximum likelihood implemented in RAxML. These individual gene trees were used to build a species tree using ASTRAL.

Descriptions of all files are listed in the file entitled README_FileExplanations.txt