Main content area

Avoiding inaccuracies in tree calibration and phylogenetic community analysis using Phylocom 4.2

Gastauer, Markus, Meira-Neto, João Augusto Alves
Ecological informatics 2013 v.15 pp. 85-90
Magnoliophyta, algorithms, competitive exclusion, fossils, phylogeny, plant communities, probability, rain forests, wildlife management
There is an increased interest in phylogenetic approaches for conservation biology and community analysis. Many of these analyses are carried out using the Phylocom 4.2 package. With this computational tool, already existing trees are pruned to species from community to be studied. For plant communities, a variety of megatrees including all angiosperm families are available for phylogenetic community analysis. Using the bladj algorithm, internal nodes of community trees derived from these megatrees are calibrated on time scales from fossil or molecular data provided in an ages file. The higher precision of tree calibration, the better is the ecological interpretation if we assume that the closest related species have the most superposed set of traits, the highest probability of co-occurrence in case of environmental filter effects and the highest ratio of competitive exclusion. Together with the Phylocom 4.2 package comes an ages file based on Wikstrom's dating of angiosperm families (wikstrom.ages). But there are inconsistencies in syntax and/or nomenclature between internal node names of trees and the ages file from phylocom that influence the tree calibration and the subsequent analysis. To avoid that, we classified all online available megatrees according to their syntax and nomenclatureof internal nodes. For each of the four classes we provide a new, fully compatible ages file in the supplement material. Each online available megatree, pruned to the species from an example community from the Atlantic Rainforest, was calibrated twice running the bladj algorithm using once the original wikstrom.ages file and, additionally, the new ages file prepared for that tree class. Outcomes from trees calibrated by different methods have been compared. To avoid inconsistencies that push results beyond the realistic, we recommend a strict application of the four ages files provided as supplementary files.