Exploring the evolution of social behaviour using genomic data

Sociality is at the root of tremendous ecological success of several taxa, including humans, ants, bees, wasps and termites. The degree and type of sociality varies greatly across taxa. The evolution of complex social behaviour can be studied by performing comparative analyses of organisms across a phylogeny showing diverse social lifestyles. We chose bees as model systems for this study because a wide range of social behaviour patterns, ranging from highly eusocial to solitary can be found in extant bees. Our aim is to identify adaptive changes in the protein coding regions of brain expressed genes. To this end, we used 454 GS FLX sequencing technology to generate the brain Expressed Sequence Tags (ESTs) of twelve socially diverse bees. The ESTs were assembled into species-specific non-redundant contigs and singletons, which were loaded into a MySQL database using custom scripts. The Honey Bee Homolog Blast website was designed to help users access the database. Users can now download these datasets or BLAST against multiple bee and wasp databases to find the homologues. The results are then sorted by e-value and displayed. The ESTs accessed through the website (http://bee12.cropsci.uiuc.edu ) can be used as a primary tool for gene discovery, genome annotation, and comparative genomic analysis.Since the Honeybee Apis mellifera had its genome recently sequenced, we designed an ortholog identification pipeline that generates multiple sequence alignments of putative orthologous genes across the twelve bees, using the gene models of Apis mellifera as the reference. The evolutionary changes associated with these alignments were then statistically inferred using maximum likelihood methods that make use of sophisticated codon-substitution models to detect non-neutral evolution in the protein coding genes. The rapidly evolving genes were then annotated using gene ontology to find over representation of associated GO terms. We also recently ventured into whole genome sequencing where we generated both single end and paired end whole genome sequence data for two of the bees, Bombus impatiens and Megachile rotundata using Illumina sequencing technology. The reads generated were assembled using a de Bruijn graph based assembly algorithm into scaffolds having a N50 of 1.12 Mb and 31 Kb respectively