Phylogenetic inference is performed using the MrBayes software and checking MCMC convergence of the tree parameter (powered by GeoKS) and evaluating post hoc the fit of the model with a posterior predictive test.

Three variants of the bayesian phylogenetic workflow exist, that differ on the method to define model of substitution:

• Automated partitioned model definition using PartitionFinder
• ​Guide the user in the choice of partitioned model using a graphic user interface
• User specified Nexus file with included full model descriptions

The phylogenetic partitioning of the diversity across samples is done using the phylogenetic entropy proposed by Chao, Chiu and Jost (2010) and equating beta diversity to mutual information between species and environment vectors. The method is implemented with a python script.

The SUPERSMART services provide functionality to infer phylogenies from a given list of taxa of interest. A set of DNA sequences for phylogenetic inference is then assembled by querying the GenBank database. To tackle the computationally challenges of inferring trees with hundreds or thousands of taxa, phylogenetic inference is accomplished in a multi-step procedure using the Maximum-Likelihood estimator implemented in ExaML and the multi-species multi-locus coalescent approach implemented in *BEAST. Computations are performed in parallel on the Naturalis biovel server.

The nexml services allow users to 1) enrich and combine datasets encoded in standard phylogenetic data formats (e.g. Newick, NEXUS, etc.) into an integrated NeXML representation and 2) to extract subsets of the data from NeXML documents.

Web services available in the BiodiversityCatalogue:

• Phylogenetics_AT_JST
• ​NeXML parser and coder
• SUPERSMART