Computational protocol: Clade Age and Diversification Rate Variation Explain Disparity in Species Richness among Water Scavenger Beetle (Hydrophilidae) Lineages

Similar protocols

Protocol publication

[…] To determine divergence times for hydrophilids we used a six-gene molecular data set of 151 species that included all major lineages of Hydrophilidae , and ran a Bayesian relaxed clock analysis with eight fossil calibrations in the program BEAST v1.7.2 . We used the following fossils to calibrate the tree (see online supplementary material for more details on fossil ages and the calibration schemes): Protochares brevipalpis (Late Jurassic, Australia) and Baissalarva hydrobioides (Early Cretaceous, Russia) Hydrobius titan (actually belonging to the genus Sperchopsis, Late Eocene, USA; Fikáček et al., unpubl. data); Limnoxenus olenus (Latest Oligocene, France) ; Anacaena paleodominica from Dominican amber (Early Miocene, Dominican Republic) , Helochares (Hydrobaticus) sp. and Cercyon sp. from Baltic amber (Eocene, Europe) (Fikáček, unpubl. data and ), and Helophorus paleosibiricus (Early Cretaceous, Russia) . We used an uncorrelated lognormal tree prior and a birth-death prior for rates of cladogenesis. The dataset was partitioned by gene with partitions unlinked and a GTR model with gamma-distributed rate heterogeneity used for each partition. We ran two analyses for 100 million generations, sampling every 1,000th generation. We verified convergence of parameter estimates and that effective sample sizes were >200 for all parameters using Tracer 1.5 . We combined runs using LogCombiner v1.6.1 and the maximum credibility tree was generated in TreeAnnotator v1.6.1 .To identify shifts in diversification rate we used MEDUSA , a comparative method that combines taxonomic and phylogenetic information to fit diversification models using stepwise addition and Akaike Information Criterion (AIC). We accounted for missing species by incorporating our species richness estimates for each major hydrophilid lineage, and pruned the tree to the most terminal clade for which species richness could be confidently estimated (see electronic supplementary material). MEDUSA uses maximum likelihood to fit birth-death, Yule, or a mixed (both birth-death and Yule) diversification models beginning with a single rate model, and using stepwise addition to add models with increasing complexity (i.e. additional rate shifts). Rate shift models are compared using AICc, with more complex models being added until the AICc threshold is no longer met and the single most likely model is selected. Due to the difficulty of estimating extinction rates from molecular data , we implemented all three options (birth-death, Yule, and mixed models).We determined the expected species richness of a clade given a net diversification rate (using background rate from MEDUSA), a relative extinction rate, and clade age using the R package Geiger . We determined the 95% confidence intervals (CI) for models incorporating high (e = 0.90) and low (e = 0.0) extinction rates. The estimated number of species for clades was plotted with the expected diversity estimates to identify clades that have significantly high or low richness given their respective ages.We used phylogenetic generalized least-squares regression and standard linear regression to test for a relationship between clade age and log-transformed species richness values using the stem clade ages (some clades were represented by a single representative preventing the use of crown ages) from our hydrophilid time tree and current figures for species richness for each major lineage compiled from literature. It is well known that species-rich lineages of insects harbor much greater diversity than is presently described . To account for this undescribed diversity, we combined our taxonomic expertise to estimate expected species richness values for each major lineage and repeated the analyses (see Table S1 in ).To explore the relative number of transitions between aquatic, semiaquatic and terrestrial habitats we conducted ancestral character reconstruction. Our taxon sampling does not allow us to determine the absolute number of transitions between these macrohabitats, but we can assess where across the entire hydrophilid tree transitions have occurred, and couple this with MEDUSA determined diversification rate shifts to explore a possible relationship between habitat type and diversification rates. We coded aquatic, semiaquatic, and terrestrial habitats as discrete, unordered character states. All character reconstructions were conducted on the maximum clade credibility tree from our BEAST analyses. We used maximum likelihood (ML) in Mesquite v2.6 to reconstruct ancestral character states under the Mk model . […]

Pipeline specifications

Software tools BEAST, GEIGER
Application Phylogenetics
Diseases Pulmonary Fibrosis