Table 1. Comparisons of the effect size* of diet on gut microbial OTUs, in different host groups (species, sex, or treatment)†.
| Comparison | Host group 1 | Host group 2 | Diet metric | ρ‡ | P§ | Difference in effect magnitude|| | Fold-difference | P§ |
|---|---|---|---|---|---|---|---|---|
| A. Between fish species† | ||||||||
| Stickleback | Perch | Proportion littoral carbon | 0.1289 | 0.0048 | s>p | 1.5 | <0.0001 | |
| Trophic position | 0.0542 | 0.2375 | s>p | 1.15 | 0.0002 | |||
| Stickleback females | Perch females | Proportion littoral carbon | 0.0988 | 0.0386 | s>p | 1.11 | 0.0005 | |
| Trophic position | 0.0900 | 0.0596 | s≤p | 1.09 | 0.0870 | |||
| Stickleback males | Perch males | Proportion littoral carbon | −0.0924 | 0.0576 | s<p | 4.7 | 0.0157 | |
| Trophic position | 0.0715 | 0.1419 | s≤p | 1.6 | 0.1727 | |||
| B. Between sexes¶ | ||||||||
| Stickleback females | Stickleback males | Proportion littoral carbon | −0.0483 | 0.2890 | f≥m | 1.29 | 0.1142 | |
| Trophic position | 0.0457 | 0.3161 | f>m | 2.25 | <0.0001 | |||
| Perch females | Perch males | Proportion littoral carbon | 0.0430 | 0.3567 | f<m | 3.26 | 0.0215 | |
| Trophic position | 0.0705 | 0.1304 | f<m | 1.89 | <0.0001 | |||
| Captive female stickleback | Captive male stickleback | Littoral versus pelagic diet | 0.1713 | 0.0077 | f>m | 1.26 | 0.0086 | |
| Captive female mice | Captive male mice | Chow versus HF diet | 0.5686 | <0.001 | m≥f | 1.14 | 0.548 | |
| Women | Men | Diet PC1 (32.1%)# | −0.0655 | 0.4785 | m≥f | 1.89 | 0.376 | |
| Women | Men | Diet PC2 (12.6%) | 0.2943 | 0.0012 | m≥f | 1.62 | 0.465 | |
| Women | Men | Diet PC3 (6%) | 0.2794 | 0.0022 | m≥f | 3.45 | 0.135 | |
| Women | Men | Diet PC4 (5.8%) | 0.1183 | 0.1999 | m≥f | 2.83 | 0.227 | |
| Women | Men | Diet PC5 (4.4%) | −0.0707 | 0.4441 | m≥f | 2.11 | 0.288 | |
| Women | Men | Diet PC6 (3.6%) | 0.1465 | 0.1117 | m≥f | 3.04 | 0.165 | |
| Women | Men | Diet PC7 (3.2%) | −0.0524 | 0.5713 | m≥f | 1.03 | 0.853 | |
| Women | Men | Diet PC8 (2.8%) | 0.1049 | 0.2559 | m≥f | 1.53 | 0.608 | |
| Women | Men | Diet PC9 (1.9%) | −0.0975 | 0.2910 | m≥f | 2.64 | 0.265 |
f, female; GLM, general linear model; m, male; OTU, operational taxonomic unit; p, perch; s, stickleback.
*The effect size of diet on each OTU was measured by the GLM-estimated slope how OTU relative abundance changes with diet, divided by the s.e. of this estimate. We obtained these effect sizes separately for all abundant OTUs within each host species (>0.01% relative abundance) and each sex within each species.
†Correlations between stickleback and perch diet effects test whether littoral/pelagic diet has a similar effect on gut microbiota of these distantly related host species. We do this both lumping sexes together, and for each sex separately.
‡Similarity in diet effects (A) between host species or (B) between sexes was assayed by testing for significant Spearman rank correlations (ρ) between OTUs’ diet effects between host groups.
§Bold denotes significant effects at P<0.05.
||Difference in average absolute magnitude of diet effects on OTUs. We used Wilcoxon signed rank tests to contrast the strength (absolute value) of diet effects between groups. We indicate which group (s, p, f, m) tends to exhibit stronger diet effects on OTUs (using ≤ or ≥ to indicate trends where nonsignificant), the ratio of mean diet effects, and a P-value.
¶Correlation between sexes within species test whether diet has similar effects on microbiota of males versus females.
#For humans, we measured diet variation using principal component axes, retaining the top nine axes that account for 72% of cumulative diet variability. Retaining nine axes was supported by a broken-stick model. For each diet axis we list the % variance explained by that axis; see Supplementary Data 3 for PCA loadings.