Master thesis defence: Bitao Qiu

The evolution of caste-specific gene expression levels in ants

Supervisors
Guojie Zhang and Jacobus J. Boomsma,  Section for Ecology and Evolution, BIO-UCPH

External examiner
Associate Professor Bent Petersen, Technical University of Denmark

Abstract
Physically differentiated queen and worker castes originated once in the common ancestor of all extant ants but caste phenotypes have later been occasionally lost or further diversified in specific subfamilies or genera. The genes responsible for caste-specific development are likely to form complex networks, so it is of interest to investigate when different components of these networks evolved and how they are differentially expressed across lineages and castes. One hypothesis suggests that genes that mediate caste-phenotypes are ancient and all derived variants of a single ancestor toolkit, while an alternative hypothesis suggests that many genes that specify adult caste phenotypes have recent origins because the environmental and social challenges of each ant lineage have been different. The few available tests of these ideas have largely focused on whole-body transcriptomes, neglecting the complex differences in transcriptomic architecture across tissues, and rarely looked at the neurological basis for behaviour, the brain. For the present study, we therefore sequenced the caste-specific brain transcriptomes across four ant species, Acromyrmex echinatior, Monomorium pharaonis, Linepithema humile and Solenopsis invicta, and inferred the origins of the genes likely to underlie their queen and worker phenotypes focusing only on genes with confirmed homology across 22 other insect species. We found that ancient genes predating the evolution of all modern insects are more likely to contribute to the regulation of gyne(queen) phenotypes and that the roles of novel genes in caste differentiation vary between species. Although examination of brain transcriptome similarity across the four-ant species revealed that subfamily-level or genus-level variation has a significant effect on overall levels of gene expression, surprisingly, we found that brain gene-expression profiles were consistently different across queen and worker castes after statistically removing these lineage-specific effects. We identified a number of pathways and genes responsible for these caste-biased expression patterns that were shared across ant species. Expression network analyses showed that conserved genetic mechanisms might be responsible for a substantial part of the variation in caste-specific gene expression in ant brains, but with significant signatures of lineage-specificity.