Characterizing larval breeding sites: bacterial community composition
In La Lopé, peridomestic larval breeding sites had a higher Shannon index than forest sites at the species, genus, and family level (Figure S4, Table S4). Other pairwise comparisons were not significant. In Rabai, Shannon index was comparable between all larval breeding site groups (Figure S5, Table S5). Ae. aegypti present sites have lower alpha diversity than Ae. aegypti absent sites at the bacterial family level, but only when we ignored habitats.
NMDS analysis suggested that forest and village (including peridomestic and domestic) larval breeding sites had a very different bacterial community in both La Lopé (Figure 2g) and Rabai (Figure 2h) at the ASV level. The forest-village divergence was less evident at higher taxonomic levels for the La Lopé larval breeding sites (Figure S6), while Rabai samples retained the difference between forest and village at all four taxonomic levels (Figure S7). Within each habitat, larval breeding sites with Ae. aegypti present and absent shared similar bacterial community composition in both La Lopé and Rabai.
When examining the most abundant bacterial families, we observed considerable variation among samples (Figure S8). Most larval breeding sites contained multiple families with no clear dominance. In La Lopé,Microbacteriaceae , Flavobacteriaceae , andBurkholderiaceae showed higher abundance in forest breeding sites, while Oxalobacteraceae and Sphingobacteriaceae were more abundant in peridomestic sites (Figure S8a, Table S6). In Rabai,Moraxellaceae had an apparent dominance in domestic sites, but its abundance was not significantly different between habitats.DESeq2 found a higher abundance of Enterobacteriaceae ,Xanthomonadaceae , Pseudomonadaceae , andPlanococcaceae in forest larval breeding sites than domestic and peridomestic sites (Figure S8b, Table S7). A full list of bacterial families that showed differential abundance between habitats are in Table S6 and S7 in the Appendix.
Lastly, NMDS analysis found that temporal samples collected from the same larval breeding sites did vary in their bacterial community composition, but the temporal changes did not exceed the variations observed within each habitat (Figure S9). That is, temporal samples from the forest clustered with the rest of forest sites instead of samples from the village, and vice versa. This result suggests that the strong divergence in bacterial communities between habitat was temporally stable.