Distinct genomic signals of lifespan and life history evolution in response to postponed reproduction and larval diet in Drosophila
Hoedjes, Katja, University of Lausanne
van den Heuvel, Joost, Wageningen University & Research
Kapun, Martin, University of Zurich
Keller, Laurent, University of Lausanne
Flatt, Thomas, University of Fribourg
Zwaan, Bas, Wageningen University & Research
Published Oct 28, 2019 on Dryad.
Cite this dataset
Hoedjes, Katja et al. (2019). Distinct genomic signals of lifespan and life history evolution in response to postponed reproduction and larval diet in Drosophila [Dataset]. Dryad. https://doi.org/10.5061/dryad.gc57610
Reproduction and diet are two major factors controlling the physiology of aging and life history, but how they interact to affect the evolution of longevity is unknown. Moreover, while studies of large-effect mutants suggest an important role of nutrient sensing pathways in regulating aging, the genetic basis of evolutionary changes in lifespan remains poorly understood. To address these questions, we analyzed the genomes of experimentally evolved Drosophila melanogaster populations subjected to a factorial combination of two selection regimes: reproductive age (early versus postponed), and diet during the larval stage (‘low’, ‘control’, ‘high’), resulting in six treatment combinations with four replicate populations each. Selection on reproductive age consistently affected lifespan, with flies from the postponed reproduction regime having evolved a longer lifespan. In contrast, larval diet affected lifespan only in early-reproducing populations: flies adapted to the ‘low’ diet lived longer than those adapted to control diet. Here we find genomic evidence for strong independent evolutionary responses to either selection regime, as well as loci that diverged in response to both regimes, thus representing genomic interactions between the two. Overall, we find that the genomic basis of longevity is largely independent of dietary adaptation. Differentiated loci were not enriched for ‘canonical’ longevity genes, suggesting that naturally occurring genic targets of selection for longevity differ qualitatively from variants found in mutant screens. Comparing our candidate loci to those from other ‘evolve-and-resequence’ studies of longevity demonstrated significant overlap among independent experiments. This suggests that the evolution of longevity, despite its presumed complex and polygenic nature, might be to some extent convergent and predictable.
This script takes count data from multiple populations in sync file format (Kofler et al. 2011) as input and calulates two-way generalized linear models with a binomial error structure for the two factors "food" and dev" and the interaction between them.
This script takes count data from multiple populations in sync file format (Kofler et al. 2011) as input and calulates (1) two-way generalized linear models with a binomial error structure, (2) two-way generalized linear models with a quasibinomial error structure, (3) generalized linear mixed models with a binomial error structure (with EE population as random factor), and (4) two-way ANOVA on arcsine square root transformed allele frequencies, for the two factors "food" and dev" and the interaction between them.
This script caluclates an empirical FDR based on permuted p-values as described in Jha et al. (2015).
Hologenome_dmel6.sh provides an overview of all the genomes (Dmel + microbes) that were obtained to generate the hologenome, which was used to map the sequence reads against. This is an updated version of the script published by Casey M. Bergman.
Reference hologenome file used in this study
European Commission, Award: FP7/2007-2011/259679
European Commission, Award: FP6 036894
Swiss National Science Foundation, Award: 310030B_176406
Swiss National Science Foundation, Award: PP00P3_165836
Swiss National Science Foundation, Award: 310030E-164207
European Research Council, Award: 741491
Deutsche Forschungsgemeinschaft, Award: Mercator Fellowship