(a) IAV-positive (IAV+) and IAV− mallards also differed in OTU richness, Shannon index, and OTU evenness results. (b) Proportions of bacterial phyla in the cloacal microbiome differed between IAV+ and IAV− mallards.
(a) Networks of cooccurrence patterns for OTUs found in mallards that tested either positive for IAV (IAV+) or negative for IAV (IAV−). Networks of D+ and D− data were constructed from the IAV+ and IAV− networks by adding edges that differed in edge weight by more than a threshold ε value of 0.2 from those of their D networks. Once D networks were created, we applied an edge threshold value of 0.5 to remove the edges with low edge weights. The clusters obtained for the D+ and D− networks with the overlapping 41 are shown next to the difference networks and are colored (in D− cluster). (b) Toy graph illustrating our method.
Log-scaled abundances of OTUs under different IAV conditions. (a) A total of 41 overlapping OTUs. (b) A total of 8 DIROM OTUs, not including the overlapping 41. (c) A total of 46 G-test OTUs, not including the overlapping 41. (d) A total of 7 OTUs that were uniquely found to be highly cooccurring in IAV+ networks. (e) A total of 80 OTUs that were uniquely found to be highly cooccurring in IAV− networks. (f) A total of 20 OTUs that were mutually significant under both IAV conditions.
(a) Venn diagram showing the degree of overlap of the following different approaches: G-test for significant differences between groups (with Bonferroni corrections for false-discovery rate), DIROM, and unique network clusters. Unique networks consist of the set of OTUs that were obtained by clustering in D+ or D− but not in both. (b) Induced subnetwork of IAV+ and IAV− for the overlapping 41 OTUs. Both subnetworks are shown as roughly complete graphs (i.e., there exists an edge within every pair of OTUs); however, the edge weights in IAV− are all of high value (average weighted degree of 25.4) and thus form a robust cluster showing strong cooccurrence patterns among the nodes, whereas in IAV+, edge weights among 41 overlapping OTUs were on average very low (average weighted degree of 1.7).
Taxonomic information for the 41 OTUs identified as contributing to observed differences in IAV− and IAV+ juvenile mallards based on the intersection of the DIROM, G-test, and network analysis resultsa
PCoA plots of IAV+ and IAV− mallards. The top row shows PCoA results calculated on log-scaled abundance values; the bottom row shows PCoA results calculated on binarized presence/absence versions of the same data. (a) A total of 10,363 OTUs found in the data after rarefaction (5,000 reads per sample). (b) A total of 10,363 OTUs minus 85 OTUs that were identified as significant by the G-test. (c) A total of 10,363 OTUs minus 47 OTUs that were identified based on the DIROM. Download FIG S1, PDF file, 0.4 MB.
A total of 85 OTUs differed significantly in mean abundance levels between IAV+ and IAV− mallards based on the G-test (P value corrected for multiple comparisons with Bonferroni method). Download TABLE S2, DOCX file, 0.1 MB.
PCoA plots of IAV+ and IAV− mallards. (a) A total of 674 OTUs found in >2 individuals per group. (b) A total of 674 OTUs minus 7 OTUs that were unique in the IAV+ network. (c) A total of 674 OTUs minus 60 OTUs that were unique in the IAV− network. (d) A total of 674 OTUs minus 20 mutual OTUs significant to both networks. (e) A total of 674 OTUs minus the 41 OTUs identified as overlapping from multiple approaches. (f) PCoA plots of IAV+ and IAV− mallards calculated on the overlapping 41 OTUs only. Download FIG S2, PDF file, 0.3 MB.