squidSim: a flexible R package for structured and reproducible simulations in Ecology and Evolutionary Biology

Joel L Pick , Hassen Allegue , Yimen Gerardo Araya Ajoy, Barbara Class , Ned A Dochtermann , Ed R. Ivimey-Cook, Kate Laskowski, Denis Réale, Jonathan Wright, Niels J Dingemanse, David F Westneat

1. Complex statistical methodology now allows a growing array of questions to be addressed in ecology and evolutionary biology. However, the particular question being addressed or the complex nature of the data collected often raise issues with how statistical models perform and potentially limit inference. Simulations provide a powerful approach to help empiricists understand the assumptions, limitations, and output of generalised linear mixed models (GLMMs), advance teaching of statistical modelling and design more informed studies around their usage.
2. Datasets in ecology and evolutionary biology often have complex hierarchical structures, which create challenges in creating simulations. This problem is exacerbated by the current lack of flexible and reproducible tools that facilitate simulating complex data from a wide range of data structures.
3. Here we present the squidSim R package, a flexible and logical program designed to accommodate many of the common data structures in ecology and evolutionary biology. The program can simulate from a wide diversity of models in a generalised linear mixed model (GLMM) framework, including data from Gaussian and non-Gaussian models, multi-response models, as well as spatial, temporal, genetic and phylogenetic effects.
4. In addition to facilitating simulations for a wide range of models and data structures, squidSim R package provides a fully reproducible workflow and has established utility for teaching. We also provide a graphical user interface via the shinySim R package.