Honey bee is an eusocial flying insect within the genus Apis of the bee clade. It’s an important pollinator, playing crucial role in both economy and terrestrial ecosystem. It also serves as a model organism for studying insect eusociality. Recent years, research on honey bee is surging, in part due to new techniques, such as next-generation sequencing, and the arrival of colony-collapse disorder (CCD), an unsolved decline in honey bee from parts of the United States, Europe, and Asia. Although the exact causes of CCD have yet to uncover, these efforts have revealed new genetic characteristics and microbial communities, improving our understanding of the evolution and health of honey bees and the challenges they face. However, so far, the understanding of the honey bee genome and metagenome is far from sufficient, and more data is urgently needed to reveal a more comprehensive figure.

Honey bee species can be classified into three distinct groups: cavity-nesting honey bees (Apis mellifera, Apis cerana, Apis koschevnikovi, Apis nulensis), giant honey bees (Apis dorsata, Apis laboriosa, Apis binghami, Apis nigrocincta), and dwarf honey bees (Apis florea, Apis andreniformis). This thesis first focused on the giant honey bees – providing the draft genome of Apis laboriosa for the first time by using whole genome shotgun sequencing, along with a comparison to its closely related species, Apis dorsata. The draft genome of Apis laboriosa based on the de novo assembly is 226.1 millions of base pairs (Mbp) in length with a scaffold N50 size of 3.34 Mbp, a GC content of 32.2%, a repeat content of 6.86%, and a gene family number of 8,404. Comparative genomics supported the close evolutionary relationship of these two species, with the evidence that Apis laboriosa showed a highly similar genomic composition with Apis dorsata, and the nucleotide divergency in single-copy genes between Apis laboriosa and Apis dorsata was only around 0.03. Interestingly, evolutionary dynamics analysis revealed that the genes in Apis laboriosa genome have undergone stronger positive selection (2.5 times more genes) and more recent duplication/loss events (6.1 times more events) than those in the Apis dorsata genome, which might be associated with the adaptation to the harsh environment in Himalayan regions. This analysis may advance understanding of the molecular basis of environmental adaptation in Apis laboriosa and provide valuable resources for future comparative/population studies of honey bee evolution.

With recent significant reductions in the cost of high throughput sequencing, metagenomics could be a useful tool for analyzing gut health and exploring the effects of environments and host on the phenotypes. To establish complete figures, this thesis secondly applied deep sequencing of the nation-wide Chinese honey bee (Apis cerana) metagenome together with a comprehensive assembly and binning analysis framework, to examine the characteristics of microbiome and their association with environments and host in different regions in China by comparing with Western honey bee (Apis mellifera) microbiome. We combined 262 newly sequenced Apis cerana metagenomic samples and 238 publicly available Apis mellifera metagenomic samples. The sampling sites of Apis cerana metagenomic samples covered 26 provinces and 14 genetically diverged populations in China, throughout a wide range of temperature zones. The sampling sites of the publicly available Apis mellifera metagenomic samples were representative as well, covering Eurasian continent. We reconstructed 728 non-low-quality Apis cerana metagenomes and 754 non-low-quality Apis mellifera metagenomes, with a high average mapping rate of 90.3% and 85.5%, respectively. These Apis cerana-derived metagenomes increase the phylogenetic diversity of bacterial genome trees by >0.08% by constructing phylogenetic tree together with 62,292 bacterial representative genomes in GTDB-R207 database and provide 109 potentially novel bacterial genomes by calculating the Average Nucleotide Identity (ANI) with public honey bee microbial reference genomes. Our analyses also reconstructed an integrative honey bee gene catalog with 1,182,294 nonredundant genes with the length > 100 bp and depth > 2. Rarefaction analysis indicated that Apis cerana metagenomes encompass fewer nonredundant genes when compared with Apis mellifera metagenomes. Similarly, overlap analysis supported that Apis cerana metagenomes contain a smaller number of unique nonredundant genes as well as functional gene contents (COG, KO, EC and Cazy). More importantly, comparison of the InrGCs with previously published honey bee gene catalogs and public protein databases revealed a large proportion of novel genes. These findings expand the characterization of the honey bee microbiota at an unprecedented resolution and greatly enlarge the current honey bee metagenome resource.

Thirdly, this thesis revealed the characteristics of both bacterial taxonomic and functional community in these two honey bee species. Comparative analysis showed that the taxonomic composition of the bacterial species is host specific. Furthermore, although Apis mellifera harbors a greater level of species richness and evenness, Apis cerana microbiome is more diverse whenconsidering the phylogenetic relationships of the microbial taxa. Also, we confirmed the overall functional profiles of the Apis cerana and Apis mellifera gut microbiomes were similar, but gut microbiota of Apis cerana is functionally less complex. The diversity analysis of Core, conditionally abundant taxa (CAT), conditionally rare and abundant taxa (CRAT), and conditionally rare taxa (CRT) in Apis cerana and Apis mellifera microbiome further revealed the nature of complexity of honey bee microbial community.

Lastly, this thesis delved into the intricate differences in the ecological assembly processes of the gut microbiomes of Apis cerana and Apis mellifera. Our results indicated that Apis cerana demonstrated a lower homogeneous selection and higher dispersal limitation and drift compared to Apis mellifera. This pattern suggested a more stochastic assembly of the Apis cerana microbiome, potentially due to higher environmental/dietary variability. Our results also revealed that non-core microbial taxa in Apis cerana demonstrated higher dispersal limitation and drift, implying a more complex and variable ecological context for these taxa, which could reflect their potential role in environmental adaptation. We finally used Variation Partitioning Analysis to disentangle and quantify the contributions of host genetics, diet, and environmental factors on the gut microbiomes of the two honey bee species. The findings indicated that diet has a more substantial influence on Apis cerana, while host genetics play a more significant role in shaping the gut microbiome of Apis mellifera. The results provided valuable insights into the co- evolutionary relationships between honey bees and their gut microbiota.

In summary, this thesis provides a clearer honey bee blueprint by comparing genomic and gut microbiome differences across species. This enhanced understanding of the characteristics of the honey bee genome and metagenome will facilitate the design of better strategies to safeguard the health of the important pollinator. Looking ahead, further research on honey bee genomics and metagenomics can deepen our understanding of insect eusociality and adaptation, contributing to biodiversity conservation, environmental health, and agricultural productivity.