Cross-Domain Ecology And Geography¶
Bijux turns reconstructed host, niche, region, and coordinate histories into reviewable branch and clade records. The owned contract is an analytical ledger with explicit uncertainty and exclusions; a map or transition count is a projection of that ledger, not an independent historical fact.
This guide owns the integration path. Use Biogeography And Migration for discrete region models and event chronology, Continuous Phylogeography And Maps for coordinate reconstruction and map ledgers, and Host Association And Niche Evolution for host and niche transitions.
flowchart LR
tree["Rooted tree<br/>with branch lengths"]
observations["Hosts · niches<br/>regions · coordinates"]
reconstruction["Discrete states or<br/>continuous locations"]
branch["Branch transitions<br/>movement and certainty"]
review["Counts · clades · rates<br/>outliers and exclusions"]
presentation["Tables · maps<br/>reproducible reports"]
tree --> reconstruction
observations --> reconstruction --> branch --> review --> presentation
Select The Historical Question¶
| Question | Native review surface | Required qualification |
|---|---|---|
| Where are host changes supported on a pathogen or parasite tree? | summarize_host_switching |
ancestral model, host coding, allowed transitions, branch certainty |
| How often and where did ecological niches change? | summarize_niche_transitions |
niche states, transition model, root treatment, clade denominator |
| Which discrete regions are connected by reconstructed movement? | summarize_discrete_region_map |
region vocabulary, centroids, state uncertainty, branch filtering |
| Which continuous movements and locations are reconstructed? | summarize_continuous_phylogeography |
coordinate units, node identity, radial uncertainty, outlier policy |
| How should an analytical movement ledger be visualized? | summarize_continuous_phylogeography_map or render_geographic_map_html |
the rendered markers and lines remain traceable to structured rows |
These surfaces summarize histories conditioned on a tree and reconstruction model. They do not identify migration, host adaptation, dispersal mechanism, or causal ecological process without additional evidence.
Host-Switching Records¶
bijux_phylogenetics.ecology:summarize_host_switching produces distinct node,
branch, count, fit, unsupported-claim, exclusion, and summary records. This
separation matters:
- node rows retain reconstructed host states and their certainty;
- branch rows retain parent and child state identity;
- count rows preserve denominators rather than reporting only selected changes;
- fit rows record model behavior;
- unsupported-claim rows prevent a convenient narrative from outrunning the reconstruction;
- exclusion rows retain taxa or branches that could not enter the analysis.
An allowed-transition table is a scientific constraint. Its revision and the host-state vocabulary belong in the run identity. A transition absent from the allowlist must not disappear; it belongs in the unsupported or exclusion record.
Ecological Niche Transitions¶
bijux_phylogenetics.ecology:summarize_niche_transitions joins discrete
ancestral states to branch transitions, counts, rates, clade summaries, and
exclusions. The report distinguishes a change in the dominant reconstructed
state from certainty about that change.
Review state order, ER/SYM/ARD parameterization, root prior, missing taxa, transition certainty class, branch length, and clade membership before interpreting a shift count. A clade with more opportunities can have more transitions without a greater rate.
Discrete And Continuous Geography¶
Discrete geography treats named regions as states. Continuous geography treats latitude and longitude as coordinates with movement distances and uncertainty. They require different contracts.
| Dimension | Discrete region model | Continuous coordinate model |
|---|---|---|
| observation | named region | latitude and longitude |
| ancestral output | state probabilities or selected state | coordinate estimate and uncertainty |
| branch output | region transition | great-circle movement and direction |
| central sensitivity | state coding and transition model | coordinate quality, covariance, and branch-time scale |
| invalid comparison | raw transition count across different vocabularies | raw distance across different units or time scales |
summarize_continuous_phylogeography emits coordinate estimates, branch
movements, robust outlier rows, exclusions, and a summary. Depth filters and
outlier thresholds are part of the analysis identity. Removing long movements
after viewing them changes the question and must create a new governed run.
Stable Node And Branch Identity¶
Node numbers from another library or tree revision are not durable identities. Use descendant-taxon signatures or another explicit structural key. Every branch row should identify its parent and child against the exact tree used for reconstruction. Tree pruning, rerooting, or topology changes invalidate position-only joins.
Integrate Without Merging Event Meanings¶
A host change, niche change, discrete region transition, and continuous coordinate movement can share a branch identity without being the same event. Join them through the exact tree and branch ledger, then retain separate model, support, exclusion, and uncertainty fields for each analysis.
| Joined observation | Valid question | Invalid shortcut |
|---|---|---|
| host switch and region transition on one branch | do independently reconstructed changes co-occur under declared support rules? | the host switch caused geographic spread |
| niche transition and coordinate movement | are inferred shifts colocated in the retained tree? | movement proves adaptation |
| discrete and continuous geography | do region-state and coordinate summaries tell a compatible bounded story? | region centroids are the inferred ancestral coordinates |
| repeated branch events across a tree set | how stable is co-occurrence across admitted trees? | the preferred tree represents event certainty |
An integrated report must preserve unmatched branches and failed component analyses. Requiring every component to succeed and then reporting only the surviving intersection changes the denominator and can manufacture apparent co-occurrence.
Presentation Is Downstream¶
render_coordinate_movement_visualization and render_geographic_map_html
consume structured marker and line records. A reviewer should be able to move
from a line on a map to its branch identity, endpoint states or coordinates,
support class, filtering decision, and source run.
Map style, colors, centroids, and viewport may change without changing the scientific record. Conversely, altering state assignments, branch inclusion, coordinate estimates, or uncertainty requires a new analytical record even if the figure looks similar.
Acceptance Record¶
Before publishing an ecological or geographic conclusion, retain:
- the tree and observation-table identities and their taxon reconciliation;
- state vocabulary or coordinate schema, units, and missing-data policy;
- reconstruction model, root treatment, transition constraints, and software identity;
- node, branch, count or movement, outlier, and exclusion rows;
- sensitivity to plausible trees, models, coding, and filters;
- a bounded conclusion that distinguishes reconstructed history from causal mechanism.
The host-switch, niche-transition, and phylogeographic writers expose separate TSV artifacts for these records. Preserve them alongside any map or narrative summary.