Biology subjects

Muff, S.

Publications and source records attributed to Muff, S..

3 recordsLinked to original sources

bioRxiv2018

Accounting for individual-specific variation in habitat-selection studies: Efficient estimation of mixed-effects models using Bayesian or frequentist computation

1. Popular frameworks for studying habitat selection include resource-selection functions (RSFs) and step-selection functions (SSFs) estimated using logistic and conditional logistic regression, respectively. Both frameworks compare environmental covariates associated with locations animals visit with environmental covariates at a set of locations assumed available to the animal. Conceptually, random coefficients could be used to accommodate inter-individual heterogeneity with either approach, but straightforward and efficient one-step procedures for fitting SSFs with random coefficients are currently lacking. 2. We take advantage of the fact that the conditional logistic regression model (i. e., the SSF) is likelihood-equivalent to a Poisson model with stratum-specific intercepts. By interpreting the intercepts as a random effect with a large (fixed) variance, inference becomes feasible with standard Bayesian techniques, but also with frequentist methods that allow one to fix the variance of a random effect. We compare this approach to other commonly applied alternatives, including random intercept-only models, and to a two-step algorithm for fitting mixed-effects models. 3. We also reinforce the need to weight available points when fitting RSFs, since models fit using "infinitely weighted logistic regression" have been shown to be equivalent to an inhomogeneous Poisson process (IPP). We generalize this result to "infinitely weighted Poisson regression", which converges to the same underlying IPP distribution. 4. Using data from Eurasian otters (Lutra lutra) and mountain goats (Oreamnos americanus), we illustrate that our models lead to valid and feasible inference. In addition, we conduct a simulation study to demonstrate the importance of including random slopes when estimating individual- and population-level habitat-selection parameters. 5. By providing coded examples using integrated nested Laplace approximations (INLA) and Template Model Builder (TMB) for Bayesian and frequentist analysis via the R packages R-INLA and glmmTMB, we hope to make efficient estimation of RSFs and SSFs with random effects accessible to anyone in the field. SSFs with individual-specific coefficients are particularly attractive since they can provide insights into movement and habitat-selection processes at fine-spatial and temporal scales, but these models had previously been very challenging to fit.

ecology↗

bioRxiv2018

Animal models with group-specific additive genetic variances: extending genetic group models

1. The animal model is a key tool in quantitative genetics and has been used extensively to estimate fundamental parameters, such as additive genetic variance, heritability, or inbreeding effects. An implicit assumption of animal models is that all founder individuals derive from a single population. This assumption is commonly violated, for instance in cross-bred livestock breeds, when an observed population receive immigrants, or when a meta-population is split into genetically differentiated subpopulations. Ignoring genetic differences among different source populations of founders may lead to biased parameter estimates, in particular for the additive genetic variance.\n\n2. To avoid such biases, genetic group models, extensions to the animal model that account for the presence of more than one genetic group, have been proposed. As a key limitation, the method to date only allows that the breeding values differ in their means, but not in their variances among the groups. Methodology previously proposed to account for group-specific variances included terms for segregation variance, which rendered the models infeasibly complex for application to most real study systems.\n\n3. Here we explain why segregation variances are often negligible when analyzing the complex polygenic traits that are frequently the focus of evolutionary ecologists and animal breeders. Based on this we suggest an extension of the animal model that permits estimation of group-specific additive genetic variances. This is achieved by employing group-specific relatedness matrices for the breeding value components attributable to different genetic groups. We derive these matrices by decomposing the full relatedness matrix via the generalized Cholesky decomposition, and by scaling the respective matrix components for each group. To this end, we propose a computationally convenient approximation for the matrix component that encodes for the Mendelian sampling variance. Although convenient, this approximation is not critical.\n\n4. Simulations and an example from an insular meta-population of house sparrows in Norway with three genetic groups illustrate that the method is successful in estimating group-specific additive genetic variances and that segregation variances are indeed negligible in the empirical example.\n\n5. Quantifying differences in additive genetic variance within and among populations is of major biological interest in ecology, evolution, and animal and plant breeding. The proposed method allows to estimate such differences for subpopulations that form a connected meta-population, which may also be useful to study temporal or spatial variation of additive genetic variance.

evolutionary biology↗

bioRxiv2018

Heritability, selection, and the response to selection in the presence of phenotypic measurement error: effects, cures, and the role of repeated measurements

Quantitative genetic analyses require extensive measurements of phenotypic traits, a task that is often not trivial, especially in wild populations. On top of instrumental measurement error, some traits may undergo transient (i.e. non-persistent) fluctuations that are biologically irrelevant for selection processes. These two sources of variability, which we denote here as measurement error in a broad sense, are possible causes for bias in the estimation of quantitative genetic parameters. We illustrate how in a continuous trait transient effects with a classical measurement error structure may bias estimates of heritability, selection gradients, and the predicted response to selection. We propose strategies to obtain unbiased estimates with the help of repeated measurements taken at an appropriate temporal scale. However, the fact that in quantitative genetic analyses repeated measurements are also used to isolate permanent environmental instead of transient effects, requires a re-assessment of the information content of repeated measurements. To do so, we propose to distinguish \"short-term\" from \"long-term\" repeats, where the former capture transient variability and the latter the permanent effects. We show how the inclusion of the corresponding variance components in quantitative genetic models yields unbiased estimates of all quantities of interest, and we illustrate the application of the method to data from a Swiss snow vole population.

evolutionary biology↗