Computational protocol: Intense natural selection preceded the invasion of new adaptive zones during the radiation of New World leaf-nosed bats

Similar protocols

Protocol publication

[…] We estimated ancestral states for 35 cranial measurements along the phylogeny using the linear parsimony approach using Mesquite version 3.02, which does not take into account the branch length when calculating ancestral values. We have also calculated ancestral states using a maximum-likelihood approach that assumes a Brownian motion model of evolution and includes the branch length in the estimates (function fastAnc of “phytools” R package). Since both methods showed equivalent patterns (correlation estimated by both methods equals 0.86, SI Appendix, Fig. ), we chose to present the results obtained by the linear parsimony ancestral reconstruction approach. After estimating the ancestral states, we calculated the response to selection vector (∆z) within each branch along the phylogeny as the difference vector between two subsequent nodes or between an extant species and its ancestor mean estimates. The net-selection gradient was then estimated based on Lande’s multivariate equation:5β=W−1Δzwhere β represents the selection gradient, ∆z is the vector of morphological change and W−1 is the inverse of the pooled within-group matrix for each node in the phylogeny.Due to the fact that inverted matrices are dominated by small eigenvalues, causing bias in the estimation of selection gradients, we used an eigenvalue extention method for noise control on the W-matrices. For this, we investigated the second derivative variance of the eigenvalues of each W-matrices. When the observed value stabilizes near zero, we replaced the eigenvalue of the subsequent eigenvectors to the stable one, using the function ExtendMatrix of the “evolqg” R package. The magnitude of selection was then estimated as the norm of the mean standardized β-vector, . [...] Each of the subfamilies of Phyllostomidae is recognized mainly due to their dietary specializations and associated morphological adaptations, , . Phyllostomidae are nested in a group of animalivorous bats, the superfamily Noctilionoidea. Basal phyllostomids kept the animalivorous habits but with a significant difference, the gleaning habit (most other noctilionoids are aerial hawking bats). Phytophagous phyllostomids, both nectar feeding and frugivorous species, evolved in clades nested among ancestral animalivorous phyllostomids, , . Previous studies of dietary diversification in phyllsotomid bats have described that the insectivorous ancestor of all phyllostomids also feed on plants, , . Dietary information was obtained from the literature based on Nowak, Ferrarezi and Gimenez, Simmons, Wetterer et al., Baker et al. and supplemented by personal experience (N. P. Giannini). The dietary information is therefore a product of our research of these sources for this paper. We organized the dietary information for the 57 species into eight categories: insectivory (includes insects and plant material in dietary habits), strict insectivory, hematophagy, omnivory, carnivory, nectarivory, frugivory and obligate frugivory. We mapped the different dietary regimes to the phylogenetic tree using the function “make.simmap” from “phytools” R package, which simulates stochastic character histories, using Markov model based on the states assigned to the tips of the tree. We used “equal rate model” because it performed better in the reconstruction compared with the other models (AIC “equal rate model”: 116,07; AIC “symmetrical rates”: 147,62; AIC “all rates different”: 191,93). […]

Pipeline specifications

Software tools Mesquite, Phytools, SIMMAP
Application Phylogenetics
Organisms Homo sapiens