Data analysis
The F and R sequences were assembled in Geneious Prime. Sequences ofOstreobium obtained from NCBI were included in the comparation
and analyses (Table 1). The alignment was performed in the MEGA X
program using the MuscleG algorithm, considering identical sequences
when performing the alignment were combined into a representative one.
The result obtained was the basis for making the selection of
evolutionary models in the jModelTest 2.1.2 software (Darriba et al.
2012), where the result was the GTR model with a Gamma distribution of
rates, according to the Bayesian Information criterion. The phylogenetic
relationships among the individuals were inferred using a Bayesian
inference (BI) method (Rambaut et al. 2018) and maximum likelihood (ML)
(Hoang et al. 2018). The outgroups were Codium fragile(AB103021), Caulerpa taxifolia (KF419327) and Caulerpa
webbiana (KY062946); in both cases the selection was based on papers
that showed the nearness of the lineage (Jackson et al. 2018; Verbruggen
et al. 2009) and authors like Gonzalez-Zapata (2018) and Gutner-Hotch
(2011) uses as base of their analyses. The ML tree topology was
estimated in IQ-Tree v.2.0 (Minh et al., 2020). Subsequently, the
Bayesian tree was useful as support to estimate divergence times. To
complain this, use BEASt v. 2.0, the substitution rate parameter assumed
a normal prior distribution whit mean= 0.00056 , SD=0.0001 (Villarreal y
Renner 2014) substitutions/site/lineage/million years. The molecular
clock model for the dataset was selected based on the likelihood radio
test in MEGA 6.0 (Tamura et al. 2013) and the result implies the use of
a uncorrelated relaxed clock, set the number of generations for the MCMC
(Markov chain Monte Carlo) analysis at 50 million and sampling every
1000 generations. Finally we used tracer to verify the quality of the
analysis and obtain parameters with an effective sample size, ESS
>200 (Rambaut et al. 2018). The substitution model
corresponded to GTR together with a gamma distribution of 4 rate
categories and the maximum credibility tree was computed in
TREEANNOTATOR (Drummond et al. 2012) removing 10% of the trees as
burn-in. The nodes indicate the posterior probability of the Bayesian
analysis, accepting a suitable grouping at probabilities ≥ 0.95. Data on
the origin of Ostreobium were based on prior studies (Del Cortona
et al. 2020; Jackson et al. 2018; Marcelino y Verbruggen 2016) with a
range of between 470 Ma and 500 Ma for its origin. To point ideas about
the grouping of the species, the GMYC (General Model Yule Coalescent)
and mPTP (multi-rate Poisson Tree Process) analyses were employed. In
this case the setting of GMYC analysis was 100000000 in MCMC, sampling
every 1000 steps and discarded 10% as burn-in. The lineages that were
congruent in both methods were considered as potential species.
The last analysis included a biogeographical history hypothesis ofOstreobium inferred using parsimony-based statistical dispersal
vicariance (S-DIVA), Bayesian binary MCMC analysis (BBM), and
dispersal–extinction–cladogenesis analysis (DEC). These three models
were implemented in RASP v. 3.0 (Yu et al. 2015). The initial parameter
was setting as default, based on the minimum result for AIC number was
made selection of analysis DEC+J, and the maximum clade credibility
(MCC) tree produced in BEAST analysis was used as the input tree. To
evaluate the degree to which Ostreobium and its hosts were
co-evolving, ML analyses of a Markov k-state parameter performed in
Mesquite v. 2.75 (Maddison and Maddison 2011) were used. Parsimony
analysis weights the contribution of each character state to a node
equally, while ML uses the information provided by the lengths of the
branches to estimate the probability that a given character state exists
at each node in the tree (Cunningham, Zhu, y Hillis 1998). The ancestral
patterns of change in the hosts were compared, paying special attention
to the lineages that showed passage between more than two hosts and the
genera that belonged to different geographic regions.
Table 1. NCBI accession codes for the sequences of rbcLgene of Ostreobium spp. present in different substrates along
locations in the world. The number of sequences was obtained after
filtering the repeat sequences and the unidentified host. The average
length was obtained after alignment.