Led description). Phylogenetic fuzzy weighting was performed in the R environment
Led description). Phylogenetic fuzzy weighting was performed in the R atmosphere (available at http:rproject.org), using the package SYNCSA .three.2 ([5], offered at http:cran.rproject.orgwebpackagesSYNCSA). Pairwise SF-837 phylobetadiversity amongst plots was obtained by computing squaredrooted BrayCurtis dissimilarities (or other acceptable resemblance measure, see Legendre Anderson [52]) for each pair of plots in matrix P (Table ). We adopted this strategy to analyze phylobetadiversity since it makes it possible for to decompose phylogenetic gradients across an array of plots into orthogonal eigenvectors and, extra importantly, to evaluate which clades are related to every single phylogenetic eigenvector [24]. We achieved this by performing a PCoA [53] depending on the squarerooted BrayCurtis dissimilarities among pairs of plots previously computed on matrix P. Such procedure generatedPLOS One particular plosone.orgprincipal coordinates of phylogenetic structure (PCPS) for every single floristic plot. Every PCPS is usually a vector describing an orthogonal phylogenetic gradient inside the dataset [8,23]. The PCPS with the highest eigenvalue describes broader phylogenetic gradients related to the split with the deepest tree nodes across the dataset, such as that connecting conifers and angiosperms. Because the eigenvalues on the other PCPS decrease, finer phylogenetic gradients related to splits of shallower nodes (e.g. families, genera) are described [8]. By relating the correlation between species from main clades and the PCPS eigenvectors, we can draw a scatterplot relating straight websites and species grouped in clades. PCPS evaluation was performed using the package PCPS (out there at http:cran.rproject.orgwebpackagesPCPS) from the R atmosphere (obtainable at http:rproject.org). Further, PubMed ID:https://www.ncbi.nlm.nih.gov/pubmed/23467991 we compared the forest kinds in relation towards the PCPS eigenvectors containing much more than five of total variation in matrix P applying oneway ANOVA. Pvalues were obtained by a permutation test with 999 iterations [37]. Such analysis allowed us to define which phylogenetic gradients had been largely associated with distinct Atlantic forest varieties. ANOVA was performed in the R environment (offered at http:rproject.org), working with package vegan two.00 ([39], available at http:cran.rproject.orgwebpackages vegan). Furthermore, we employed other four wellknown phylobetadiversity measures to compare the forest types inside the Southern Brazilian Atlantic Forest (see Table ). COMDIST is usually a phylobetadiversity measure that computes the mean phylogenetic distance among species occurring in two distinctive web pages [44]. For this reason, this phylobetadiversity measure captures variation associated with the extra basal nodes linking species [3]. Computing COMDIST values without having thinking of the variation in species abundances is equivalent to compute the phylogenetic distinctness (Rao’s D) proposed by Hardy Senterre [50]. Hence, we opted for employing only the former in this study. Alternatively, by standardizing Rao’s D values by the mean withinsite phylogenetic diversity it’s probable to acquire yet another phylobetadiversity measure (Rao’s H, [50]), which captures phylobetadiversity patterns related to more terminal nodes inside the tree [3]. COMDISTNT [44] measures the imply phylogenetic distance among each and every species within a plot and also the nearest phylogenetic neighbor in a different website (Table ). It can be, therefore, a “terminal node” metric [3]. The last phylobetadiversity method used within this study was UniFrac [49], which measures, for every pair of websites, the fraction.