Computational protocol: Change in Phylogenetic Community Structure during Succession of Traditionally Managed Tropical Rainforest in Southwest China

Similar protocols

Protocol publication

[…] We constructed a phylogenetic tree (, ) with the APGIII classification as a backbone for all the species in our study site using the program PHYLOMATIC v 4.0.1 . However, this tree was only resolved up to family level. To increase the resolution of this phylogeny we recovered rbcL (548–697 bp) sequences for 163 species from GenBank. For another 76 species in our dataset, for which no rbcL data was available, we assigned the rbcL sequence of the closest available congener. This is a reasonable approach since rbcL is a slowly evolving sequence and shows very little variation below the genus level , . Five species for which we found no rbcL sequences and which also had no congeneric relative with known rbcL sequence were excluded from the analysis. These five species were rare in the sampled communities, and should not bias our further analyses, especially as these individuals were not clustered into a particular phylogenetic lineage. Using the rbcL sequences, one most likely phylogenetic tree was built under the GTR model of evolution with a maximum likelihood approach using the software PAUP version 4 with 100 bootstrap replicates to assess node support. The phylogenetic hypothesis built with the rbcL gene was broadly consistent with the topology of the APG III phylogeny. In the rbcL tree, most families had a bootstrap node support >80. Below the family level, most of the genera appeared well resolved although with lower bootstrap support values. We then used the APGIII phylogeny as generated by PHYLOMATIC as our family level backbone phylogeny and resolved the within family phylogenetic classification by adding the results obtained from the rbcL phylogeny. Branch lengths and dated nodes for this megatree were obtained by applying the bladj algorithm of PHYLOCOM , with calibration ages from . [...] In order to standardize plot comparisons all diversity analyses described below were based on a fixed number (based on the plot with the fewest stems) of randomly selected individuals from each plot. This rarefaction procedure reduces the impact of sample size differences between plots on the diversity analyses . We calculated phylogenetic species variability (PSV), a measure of the phylogenetic relatedness of species within a community, of each plot based on species presence/absence and their phylogenetic relationships within the rarefied samples . Faith’s phylogenetic diversity (PD) of each plot was calculated as the sum of branch lengths of the subtending tree of the species present in the rarefied samples. We calculated phylogenetic species richness (PSR) for each plot by multiplying the number of species in the rarefied samples by their PSV value, and calculated phylogenetic species evenness (PSE) by incorporating relative species abundances within the rarefied samples into PSV . All these phylogenetic diversity indices were calculated by using PHYLOCOM and the R package picante .The phylogenetic similarity of the plant communities was assessed using Unifrac . UniFrac estimates the distance between communities as the fraction of the branch length of the phylogenetic tree that leads to descendants from either one environment or another, but not both. We used the resulting distance matrix to cluster environments using Jackknifed UPGMAwith R package picante .We calculated the abundance weighted net relatedness index (NRI) and nearest taxon index (NTI) of each plot to measure the phylogenetic dispersion (relatedness) of the co-occurring species by using PHYLOCOM . The NRI and NTI for each rarified sample were calculated as:(1)(2)Where MPD is the mean pairwise phylogenetic distance between all individuals in each rarified sample and MNTD is the mean phylogenetic distance for each individual to its nearest relative within each rarified sample. The MPDrandom and MNTDrandom are the mean MPD and mean MNTD from 999 randomly generated assemblages. An independent swap null model was used to generate these 999 random assemblages. For both NRI and NTI, values close to zero indicate random phylogenetic structures, positive values indicate clustered, while negative values indicate overdispersed community phylogenetic structures. […]

Pipeline specifications

Software tools Phylomatic, PAUP*, Phylocom, Picante, UniFrac
Applications Phylogenetics, 16S rRNA-seq analysis
Organisms Homo sapiens