1. Distinguish this route’s prerequisite from the general concept
A tanglegram places two phylogenies side by side and connects corresponding taxa. This app’s Tanglegram route uses a particular ordering rule: the left order is the taxon-index order, and the right order sorts the row sums of a sequence-derived distance matrix. It requires that matrix to have one row and one column per taxon.
That distance-matrix requirement belongs to this implementation. It is not a universal requirement for comparing two trees with a tanglegram. Supplying two valid Newick trees can satisfy the app’s general input loader while leaving this selected method without its required matrix.
The reviewed source starts with all twenty methods selected. The retained calls enabled only tanglegram_; the saved output is named Tanglegram Layout. Each case reached this route once and reached no other analysis route. The inputs were supplied through the public method contract, not a demonstrated Load dialog or hosted upload.
2. Read the complete Newick-only refusal
The first case supplied two named six-taxon trees, First and Second, with the same exact text:
((A,B),C,(D,(E,F)));
The input also retained nj_ values [10, 20, 30] and upgma_ values [10, 30, 20]. Those arrays belong to a different input contract; they do not provide a taxon-indexed distance matrix. No alignment was supplied in this case.
The public input path loaded the trees, but the selected route received an empty one-dimensional matrix with shape [0]. It returned false before creating crossing statistics and recorded all three fields below:
| Saved field | Saved value |
|---|---|
| Status | no_data |
| Required upload | an uploaded alignment yielding a taxa-indexed distance matrix |
| Ref | Scornavacca et al., J Graph Alg Appl 2011 |
The app’s has_ flag remained true and the overall log ended with Complete. The selected method still had no_. Its saved internal result event carries the same requirement. A loaded-input flag or completion message does not turn a refusal into a zero crossing count.
Source inspection shows that two valid supplied trees take precedence over the sequence branch. These retained cases test trees and sequences separately; they do not establish mixed-input behavior. Preserve the original inputs and requirement text rather than adding a fabricated distance matrix.
3. Preserve the exact positive alignment and row order
The positive case supplied four equal-length synthetic sequences through the sequences key, in this exact order. Each row has twelve bases:
| Taxon | Saved sequence |
|---|---|
| A | ACCAAAAAAAAA |
| B | AAAAAAAAAAAA |
| C | AAACCCCAAAAA |
| D | CAAAAAAAAAAA |
The saved public input contains complete four-by-four proportion-difference and corrected-distance matrices. The original distance, NJ and UPGMA helpers were reached. That reachability does not independently validate the inferred tree topologies or their branch lengths; those methods need their own scientific evidence.
Keep these names in A, B, C, D order when identifying the retained example. The source builds matrices in input order and later sorts taxon labels. The saved fixture already has sorted labels, so it does not establish correct row-label association for reordered records. Tied row sums, alternative taxon orders, gaps, ambiguous or missing bases, saturated distances and larger alignments remain outside these two cases.
4. Interpret the observed orders and crossings
The retained left order is A, B, C, D. The right order is B, D, A, C, obtained from the ascending distance-row sums. The source calls these orders “NJ” and “UPGMA,” but this ordering step does not traverse the two inferred trees to optimize their leaf embeddings.
| Position, starting at 1 | Left taxon | Right taxon |
|---|---|---|
| 1 | A | B |
| 2 | B | D |
| 3 | C | A |
| 4 | D | C |
The earlier independent reference records three crossing pairs: A–B, A–D and C–D. It records a maximum of six pairs and a crossing fraction of 1 / 2. The saved method values are 3, 6 and 0.5. These are copied from the retained reference and observation; no distances, orders or crossings were recalculated for this lesson.
The complete successful summary contains five fields:
| Saved field | Saved value |
|---|---|
| Crossings | 3 |
| Max Crossings | 6 |
| Crossing Fraction | 0.5000 |
| Quality | Poor |
| Ref | Scornavacca et al., J Graph Alg Appl 2011 |
The fraction is formatted to four decimal places before the table stores it. Poor is the source’s label when the fraction is not below 0.3; it is not a calibrated biological quality judgment. A fraction of one half does not mean 50% support for either tree or 50% evolutionary disagreement.
5. Separate the ordering rule from the cited method
Scornavacca, Zickmann and Huson’s tanglegram paper describes an NN-tanglegram approach for rooted trees and networks. Its bibliographic record is Bioinformatics 27(13), i248–i256 (2011). The saved table’s “J Graph Alg Appl 2011” reference string is preserved above, but it does not match that record.
This source’s distance-row-sum ranking does not establish implementation of the paper’s NN-tanglegram algorithm or a minimum-crossing tree layout. The retained controls verify only the recorded prerequisite and one ordering calculation. They do not establish topology-preserving embeddings, general optimization or scientific agreement with that publication.
Graphics were disabled in both retained cases. No native tanglegram or crossing-distribution plot was drawn. The source’s optional histogram uses cyclic shifts; these checks do not validate that plot as a sampling distribution or a significance test.
6. Retain evidence without inventing an export
The observations retain the exact named inputs, both complete tables, two full distance matrices, orders, logs and refusal event. The earlier independent Decimal reference is retained separately from the binary floating-point matrices. A greater number of reference digits does not add precision to the application’s saved numbers.
These two cases produced zero native exports. The website evidence JSON is a written-guide attachment, not a verified Save Table or Export All package. Loading, native viewer feedback, figures, CSV completeness, saved sessions and hosted delivery require separate evidence.