Fungi in the marine environment are often neglected, despite that they having critical roles on land as decomposers, pathogens or beneficial endophytes. We used culture-dependent methods to survey the fungi associated with the seagrass, Zostera marina, and then placed these sequences in the fungal tree of life. Four different sequence alignments were generated, (1) an alignment to investigate seagrass isolates in the Basidiomycota and Zygomycota phyla, (2) an alignment to investigate seagrass isolates in the Eurotiomycetes class in the Ascomycota phylum, (3) an alignment to investigate seagrass isolates in the Sordariomycetes class in the Ascomycota phylum, and (4) an alignment to investigate seagrass isolates in the Dothideomycetes class in the Ascomycota phylum. 

Each of the four sequence alignments was generated using MAFFT (v. 7.402) (Katoh et al., 2002) with default parameters on the CIPRES Science Gateway web server and trimmed using trimAl (v.1.2) with the -gappyout method (Capella-Gutierrez et al., 2009). Sequence alignments were then further trimmed to the D1/D2 regions of the 28S rRNA gene with trimAl using the select option (e.g. Basidiomycota / Zygomycota alignment { 614-2899 }, Eurotiomycetes alignment { 0-569 }, Sordariomycetes alignment { 501-1224 }, and Dothideomycetes alignment { 0-429, 993-1755 }). Spurious sequences (e.g. sequences which contained few or no nucleotides after trimming) were then removed with trimAl using -resoverlap .75 -seqoverlap 50. The resulting alignments contained: 80 sequences with 614 positions (Basidiomycota / Zygomycota), 91 sequences with 509 positions (Eurotiomycetes), 96 sequences with 501 positions (Sordariomycetes), and 107 sequences with 563 positions (Dothideomycetes).

JModelTest2 (v. 2.1.10) was run and the best-fit model for all alignments was the GTR + I + G evolutionary model. Bayesian phylogenetic inference for each alignment was performed using MrBayes (v. 3.2.2) with four incrementally heated simultaneous Monte Carlo Markov Chains (MCMC) run over 10,000,000 generations. The analysis stopped early if the optimal number of generations to reach a stop value of 0.01 or less for the convergence diagnostic was achieved (Huelsenbeck & Ronquist, 2001). This occurred for the Eurotiomycetes, Sordariomycetes and Dothideomycetes alignments at 2,150,000 generations, 1,375,000 generations and 2,140,000 generations, respectively. The Basidiomycota / Zygomycota alignment ran for the full 10,000,000 generations, only achieving an average standard deviation of split frequencies of 0.049. The first 25% of trees generated for each alignment were discarded as burn-in and for the remaining trees, a majority rule consensus tree was generated and used to calculate the Bayesian Posterior Probabilities.

Here we provide the trimmed alignment files used to generate all four phylogenetic trees and the four phylogenies themselves.