posted on 2023-11-20, 05:20authored byJavad Sadeghi, Subba Rao Chaganti, Timothy B. Johnson, Daniel D. Heath
Additional file 1: Figure S1. Relative abundance of bacterial community composition presented at the phylum level for gut, skin, and water microbiota (samples are combined across sample types). Phyla with less than 0.01% of relative abundance are combined and presented as “others”). Figure S2. Bacterial community composition (relative abundance at the family level) for gut, skin, and water microbiomes across all fish species collected at three sites in the Great Lakes (Lake Erie, Lake Ontario and Detroit River). Bacterial families with less than 0.1% relative abundance are combined and presented as “others”. Fig S3. LASSO regression analysis was used to indentify the best predictor variables for gut samples based on their alpha (Shannon entropy, PD, Chao1) and beta diversity (PCoA 1- 5) indices. Inside each bar is showing the coeffitient value and X axis is showing the importance of the predictor variable for alpha and beta diversity indices. Diet, location and fish species was identified as the best predictor for most of the diversity indecies. Fig S4. LASSO regression analysis was used to indentify the best predictor variables for fish skin samples based on their alpha (Shannon entropy, PD, Chao1) and beta diversity (PCoA 1- 5) indices. Inside each bar is showing the coeffitient value and X axis is showing the importance of the predictor variable for alpha and beta diversity indices. Location and fish species was identified as the best predictor for most of the diversity indecies. Figure S5. Scatterplot of pairwise host phylogenetic distance vs pairwise Bray Curtis dissimilarity for both gut (a) and skin (b) samples. Samples were combined within host species. Host phylogenetic distance was estimated using of CO1 and CytB mitochondrial gene sequences. Table S1. Summary of Great Lakes fish species sampled for gut and skin microbiome. We provide a description of the sample locations and total sample size. Table S2. Comparison of differentially abundant bacterial taxa at the family level for gut and skin microbiomes across all fish species and sample locations using DESeq2 method (Benjamini-Hochberg false-discovery rate [BH FDR] 0.05, |log2fold change| > 2). Positive log2 FC indicate higher abundance in skin samples and negative log2 FC specify higher abundance in gut samples.
Funding
Natural Sciences and Engineering Research Council of Canada