10.5061/DRYAD.BS41Q
Simon, Monique Nouailhetas
University of Sao Paulo
Marroig, Gabriel
University of Sao Paulo
Data from: Evolution of a complex phenotype with biphasic ontogeny:
contribution of development versus function and climatic variation to
skull modularity in toads
Dryad
dataset
2018
modularity models
P-matrices
Rhinella granulosa complex
2018-10-13T00:00:00Z
en
https://doi.org/10.1002/ece3.3592
84616 bytes
1
CC0 1.0 Universal (CC0 1.0) Public Domain Dedication
The theory of morphological integration and modularity predicts that if
functional correlations among traits are relevant to mean population
fitness, the genetic basis of development will be molded by stabilizing
selection to match functional patterns. Yet, how much functional
interactions actually shape the fitness landscape is still an open
question. We used the anuran skull as a model of a complex phenotype for
which we can separate developmental and functional modularity. We
hypothesized that functional modularity associated to functional demands
of the adult skull would overcome developmental modularity associated to
bone origin at the larval phase because metamorphosis would erase the
developmental signal. We tested this hypothesis in toad species of the
Rhinella granulosa complex using species phenotypic correlation pattern
(P-matrices). Given that the toad species are distributed in very distinct
habitats and the skull has important functions related to climatic
conditions, we also hypothesized that differences in skull trait
covariance pattern are associated to differences in climatic variables
among species. Functional and hormonal-regulated modules are more
conspicuous than developmental modules only when size variation is
retained on species P-matrices. Without size variation, there is a clear
modularity signal of developmental units, but most species have the
functional model as the best supported by empirical data without
allometric size variation. Closely related toad species have more similar
climatic niches and P-matrices than distantly related species, suggesting
phylogenetic niche conservatism. We infer that the modularity signal due
to embryonic origin of bones, which happens early in ontogeny, is blurred
by the process of growth that occurs later in ontogeny. We suggest that
the species differing in the preferred modularity model have different
demands on the orbital functional unit and that species contrasting in
climate are subjected to divergent patterns of natural selection
associated to neurocranial allometry and T3 hormone regulation.
Phenotypic matrices with size variationPearson product-moment correlation
matrices of all toad species for the 21 linear
distances.P_matrices_ECE-2017-06-00822.csvList of covariance matrices of
all toad speciesList of residual covariance matrices for the 21 linear
distances extracted from toad skull. The matrices were constructed using
residuals of multivariate linear models controlling for sex and locality.
This list can be read in R using load('cov.list_data'). To
construct correlation matrices, use lapply(cov.list,
cov2cor).cov.list_dataDefinition of each developmental, hormonal and
functional unitData frame describing which distances belong to each
developmental, hormonal or functional unit. 1 for distances that belong to
the unit and 0 for distances that do not belong to the unit. May be used
as the argument 'modularity.hypot' in the function
'TestModularity' of the 'evolqg' R
package.mod_hypo.csvModularity models for P-matrices with size
variationData frame describing the modularity models used to test
P-matrices with size variation. May be used as argument 'mod' in
the 'EMMLi' function in R.mod_models_size.csvModularity models
for P-matrices without allometryDescription of modularity models used to
test P-matrices without allometric size variation. May be used as argument
'mod' in 'EMMLi' R
function.mod_models_nosize.csvModularity models for P-matrices without
isometryDescription of modularity models used to test P-matrices without
isometric size variation. May be used as argument 'mod' in
'EMMLi' R function.mod_models_noiso.csv