The influence of environment on mosquito feeding patterns: a meta-analysis of "universal" DNA diet studies in a global context

Mosquitoes have innate preferences for their blood-meal hosts, but these can be modified by the environment, with implications for disease spread under climate and land use change. To predict the spread of vector-borne pathogens more accurately we need to better understand blood-meal plasticity under changing environmental conditions. We compiled blood-meal studies for six prominent disease vectoring mosquitoes from around the globe (Aedes aegypti, Ae. albopictus, Anopheles funestus, An. gambiae, Cx. pipiens, and Cx. quinquefasciatus). We targeted studies that used universal molecular methods (typically PCR/metabarcoding) to identify hosts from a wide range of candidates (as opposed to studies using methods that presuppose host identity - such as precipitin, ELISA). We found that blood-meals from the >15,600 analysed were mostly from the expected host-groups for each mosquito species, but we frequently encountered atypical hosts (e.g. mammalophilic species feeding on birds/reptiles). The universal methods used by the studies in our metanaalysis identified high host richness, and we found [≥]174 hosts for Culex and [≤]65 species for Aedes mosquitoes at a considerably increased discovery rate of novel hosts per sampling effort. We used a hierarchical Dirichlet regression model to analyse global variation in feeding patterns in relation to environmental datasets (land use, precipitation, mean annual temperature, latitude, human and livestock density). Land use, mean annual temperature and poultry density had noticeable effects on blood-meal selection of Ae. aegypti, Cx. pipiens and Cx. quinquefasciatus. Although host density was a factor in blood-meal selection - host choice is not a simple function of host availability, as has previously been observed, but contingent on other drivers. While our compiled dataset afforded us these insights, improving resolution and consistency of data gathering and reporting would improve the precision of how blood-meal studies can inform us of present and potential risks of pathogen transmission events.