Abstract
Mosquitoes harbor diverse microbial communities that influence their potential to transmit pathogens. However, the ecological drivers shaping these microbiomes, particularly in under-sampled regions like Africa, remain poorly resolved. We conducted a large-scale metatranscriptomic survey of 3,940 Aedes and Culex mosquitoes from diverse ecological zones across Kenya. Our analyses revealed that viruses dominated the overall transcriptome, while bacteria exhibited the greatest taxonomic richness. Geographic location emerged as the primary driver of microbial community structure, whereas host genus identity shaped virome diversity at local or city-level scales. Culex mosquitoes harbored higher viral richness, particularly in coastal regions, while Aedes supported more diverse bacterial assemblages. Microbial co-occurrence networks exhibited distinct topologies across hosts: Culex networks featured cross-domain interactions and viral keystone taxa, whereas Aedes networks were more cohesive and robust, centered on bacterial hubs. We identified 102 distinct viruses from 24 families, including 31 putative novel RNA viruses. Segment-resolved phylogenies revealed cryptic clades within Bunyavirales, Picornavirales, and other lineages. Collectively, our findings highlight the scale-dependent influences of geography and host identity on mosquito microbiomes in East Africa and demonstrate the utility of metatranscriptomics in uncovering hidden microbial diversity and ecological interactions. These insights provide a foundation for ecologically informed arthropod vector surveillance and microbiome-based intervention strategies.IMPORTANCEMosquitoes are more than just flying syringes; they are complex ecosystems hosting a variety of microbes. Understanding what shapes this microbial world inside mosquitoes is key to developing new control strategies. Our study of nearly 4,000 mosquitoes from Kenya reveals that where a mosquito lives matters most for its overall microbial makeup, but its genus dictates which viruses it carries. We discovered that different mosquito types have distinct microbial social networks: one type has a fragile network centered on viruses, while the other has a resilient network built around bacteria. This means that strategies to disrupt disease transmission by targeting mosquito microbes may need to be tailored to a specific mosquito genus. Our work provides a map of these microbial ecosystems, highlighting potential new viruses and offering insights for future public health surveillance and interventions.