全球变化知识资源中心

Analyses in many fields of ecology are increasingly considering multiple species and multiple individuals per species. Premises of statistical tests are often violated with such datasets because of the non-independence of residuals due to phylogenetic relationships or intraspecific population structure. If comparative approaches that account for the phylogenetic relationships of species are well developed and their benefits demonstrated, the importance of considering the intraspecific genetic structure, especially in combination with the phylogenetic structure, has rarely been addressed. We investigated whether it is beneficial to account for intraspecific genomic relatedness in multi-species studies. For this, we used a phylogenetic mixed model to analyse first a suite of simulated data and then results from one example ecological study-a budburst experiment where clippings of 10 tree and shrub species were subjected to different temperatures and photoperiods. We found that accounting for intraspecific genetic structure yields more accurate and precise fixed effects as well as increased statistical power, but more so when the relative importance of the intraspecific to the phylogenetic genetic structure is greater. Analysis of the budburst experiment further showed that accounting for intraspecific and phylogenetic structures yields improved estimates of warming and photoperiod effects and their interaction in explaining the time to budburst. Our results show that statistical gains can be made by incorporating information on the intraspecific genomic relatedness of individuals in multi-species studies. This is relevant to investigations that are interested in intraspecific variation and that plan to include such observations in statistical tests.