Rabies Comparative Report
This comparative section asks whether the host-associated lineages in the rabies demonstration tree are associated with a consistent eastward geographic placement when geography is summarized as regional longitude. The result is interpretive evidence, not causal proof, and it inherits the small-panel limits of this dataset.
formularegion_longitude ~ host_group
analysis taxa9
selected trait modelbrownian
pgls r-squared0.833944827574
Question and Answer
- Question: does host association coincide with a consistent longitudinal shift in this rabies panel?
- Answer: `host_group[canid]` shows a positive association and is nominally supported.
- Phylogenetic signal remains strong for the response trait (Blomberg's K 2.44100950126, Pagel's lambda 1).
- Interpret the coefficient evidence cautiously because the residual diagnostics retain review warnings and the sample is intentionally compact.
Coefficient Summary
| term | estimate | standard_error | p_value | significant |
|---|---|---|---|---|
| intercept | -58.440449865 | 0.119088670303 | 4.88498130835e-15 | true |
| host_group[canid] | 144.746258822 | 52.8581368031 | 0.0338077145602 | true |
| host_group[livestock] | 154.578389271 | 55.4459300824 | 0.0316646120912 | true |
Residual Diagnostics
| analysis | residual_variance | max_abs_standardized_residual | phylogenetic_residual_lambda | warnings |
|---|---|---|---|---|
| brownian | 4977.69841111 | 0.964746416527 | 1 | residuals retain moderate phylogenetic structure |
| ou | 4977.69841111 | 0.952510163321 | 1 | residuals retain moderate phylogenetic structure; best-supported OU alpha falls on the search boundary and may not be well identified; OU likelihood surface is shallow across alpha values, so model choice may be unstable; best-supported OU alpha is weak and may be difficult to distinguish from Brownian motion |
| pgls | 13904.1866087 | 4.29047006893 | 1 | PGLS residuals retain moderate phylogenetic structure; PGLS residual diagnostics identify one or more outlier taxa; PGLS leverage diagnostics identify one or more high-leverage taxa |
Comparative Tree Adjustments
- The comparative fit uses the rooted demonstration tree after flooring any nonpositive nonroot branch lengths to a tiny positive value.
- Adjusted branch count: 1
| node_label | original_branch_length | repaired_branch_length | reason |
|---|---|---|---|
| bat_chile_rv108 | 0 | 1e-06 | comparative branch-length methods require strictly positive nonroot branch lengths |
Interpretation Ledger
| topic | claim | evidence | caution |
|---|---|---|---|
| formula | comparative regression was fit as `region_longitude ~ host_group` | analysis_taxa=9; predictors=1 | formula interpretation is conditional on the supplied tree, taxon overlap, and encoded predictor structure |
| phylogenetic-signal | the response trait retains high phylogenetic structure | blombergs_k=2.441010; pagels_lambda=1.000000 | signal metrics describe covariance with the supplied phylogeny and are not by themselves a process explanation |
| model-comparison | brownian is the preferred continuous trait model on AICc | aicc_delta=4.934486; selected_from=2 models | small AICc differences should not be overstated as decisive biological process separation |
| coefficient | `host_group[canid]` shows a positive association and is nominally supported | estimate=144.746259; p_value=0.033808; 95%_ci=[15.407057, 274.085460] | coefficient interpretation remains associational and depends on lambda choice, sample size, and residual diagnostics |
| coefficient | `host_group[livestock]` shows a positive association and is nominally supported | estimate=154.578389; p_value=0.031665; 95%_ci=[18.907086, 290.249693] | coefficient interpretation remains associational and depends on lambda choice, sample size, and residual diagnostics |
| diagnostics | brownian residual diagnostics flagged review concerns | residuals retain moderate phylogenetic structure | diagnostic warnings weaken confidence in coefficient interpretation and should be reviewed before publication use |
| diagnostics | ou residual diagnostics flagged review concerns | residuals retain moderate phylogenetic structure; best-supported OU alpha falls on the search boundary and may not be well identified; OU likelihood surface is shallow across alpha values, so model choice may be unstable; best-supported OU alpha is weak and may be difficult to distinguish from Brownian motion | diagnostic warnings weaken confidence in coefficient interpretation and should be reviewed before publication use |
| diagnostics | pgls residual diagnostics flagged review concerns | PGLS residuals retain moderate phylogenetic structure; PGLS residual diagnostics identify one or more outlier taxa; PGLS leverage diagnostics identify one or more high-leverage taxa | diagnostic warnings weaken confidence in coefficient interpretation and should be reviewed before publication use |