BF Workbench is a restricted scientific system. Records are available only to registered, authorised accounts.

Tree construction

Alignment, tree-building programs and output styles.

Reference

Tree Construction builds a tree from the sequences of the selected records for one marker and draws it as a multi-page PDF. Use it to see how sequences cluster: whether the records of one species group together, or where a case sequence falls among reference records. It is illustrative: it does not root the tree and, with the exceptions noted below, attaches no support values.

Purpose

The question answered is "how do these sequences group when a tree is built from them?". The method makes no identification and no species delimitation; for those, see Identification and Barcode gap.

Inputs

Records selected on the console, restricted to one marker. The shared record filter (see The methods and their shared pipeline) keeps a record when its marker_code matches, its count of A, C, G and T bases lies within the length bounds, and no more than 10% of its bases are ambiguity codes. A record with no sequence is dropped. Each surviving record becomes one tip named by its processid. The fields chosen as branch labels, and phylum, are carried through for labelling.

Parameters

Parameter Form offers Job applies
Tree Type Multipage Phenogram or Circular Phylogram Not applied. The drawing is always a rectangular tree, 100 tips per page.
Tree Method QuickTree, FastTree, RAxML, IQ-TREE Applied by program name. The model hint in the option value (for example [p] or [gtrgamma]) is not read.
Marker one marker Applied as the marker_code filter
Align Sequences None or MAFFT Applied
Branch labels: specimen, taxonomy, collection, sequence fields from each group Applied. Values are joined with \| into the tip label.
Branch labels: record organization, forensic extension, forensic extension advanced, feature rollup fields from each group Not applied
Sequence Length Filter 400–1000 by default Applied, counting A, C, G and T only
Exclude Records with Stop Codons checkbox Not applied
Exclude Records Flagged as Misidentifications or Errors checkbox Not applied
Colorize Tree Based on phylum, class, order, family, subfamily, tribe, country, province, BIN or none Not applied to colours, which are always by phylum. The chosen field is appended to each tip label.
Ambiguous Base/Gap Handling pairwise or complete deletion Not applied. A value is required, but the alignment reaches the tree program unchanged.
Minimum Complete Overlap 50–1000, default 100 Not applied

The report lists every parameter as submitted, including those the job does not apply.

Method

  1. Alignment. With MAFFT, the job runs MAFFT in automatic mode (--auto), which chooses the strategy from the size of the data set. With None, the sequences pass through as stored, and nothing checks that they are the same length.
  2. Input format. QuickTree reads Stockholm format; the converter pads any shorter sequence with trailing gaps to the length of the longest. The other programs read the FASTA directly.
  3. Tree building. One program runs, chosen by the Tree Method: - QuickTree: neighbour joining on Kimura-corrected distances (-kimura). The option value quicktree[p] suggests p-distance, but the job always applies the Kimura correction. - FastTree: approximate maximum likelihood under the GTR nucleotide model (-gtr -nt), in fastest mode (-fastest) with the second round of nearest-neighbour interchanges skipped (-no2nd), on 4 threads. - RAxML: raxmlHPC maximum-likelihood search under GTRCAT from a parsimony starting tree with the fixed seed 12345; the best tree is kept. - IQ-TREE 2: maximum likelihood under GTR+G with the fast search (-fast) and an SH-aLRT test of 1000 replicates, on 4 threads. No random seed is set.
  4. Drawing. The tree is rendered to PostScript with newicktops from njplot at 100 tips per page, and each tip's process ID is replaced with its full label in colour. The palette has 20 colours, assigned to phyla in order of first appearance and repeated after the twentieth. Ghostscript converts the pages to PDF.
  5. Packaging. The standard report is generated and the tree pages are appended to it.

Outputs

File Content
tree.nwk Newick tree, tips named by processid, with branch lengths. IQ-TREE writes SH-aLRT values on internal nodes; FastTree writes its default local support values; QuickTree and RAxML write none.
output.pdf The drawn tree, 100 tips per page
analysis_report.pdf, analysis_report.html The report (no charts for this method); the PDF has the tree pages appended
filtered_records.jsonl The records that became tips

The package also holds the working files: the FASTA before and after alignment, the program's own output files (for example iqtree_run.* or RAxML_*), the PostScript pages, and a one-line text file listing the label columns, which carries no result data. See Anatomy of a result package.

Limitations and interpretation

  • The tree is unrooted. The drawing places a root arbitrarily; do not read direction or ancestry from it.
  • No bootstrap for any method. The only support values are IQ-TREE's SH-aLRT and FastTree's local supports. RAxML and QuickTree trees carry none.
  • One style only. Circular layouts, cladograms and unrooted layouts are not produced, whatever the form says.
  • Alignment quality is yours to judge. The job records no alignment statistics. With None and QuickTree, unaligned sequences are padded, not aligned, and the distances are not meaningful.
  • Reproducibility. IQ-TREE runs without a fixed seed, so two runs on the same data can differ in topology or support.
  • Colouring. Colours mark phylum only; a single-phylum data set is drawn in one colour. The field chosen under Colorize Tree Based on still appears at the end of each label.
  • Tool versions are collected on Software, versions and citations; of the tree programs, only IQ-TREE's version is pinned.