GitHub
08/31/2026, 5:53 PMtestdata/phylosearch_input_reftree.csv for nf-core/phyloplace#78 — a two-row "search and place" sample sheet whose rows share a reftreename and are therefore summarised jointly as well as individually.
Follows on from #2250, which added the column to phyloplace_input.csv.
Why a separate file
phylosearch_input.csv is deliberately left untouched, so it keeps covering the layout of a sample sheet written before the reftreename column existed. Between the two files the suite then tests both layouts in that mode: without the column, and with a group.
What is in it
Two rows, both searching testdata/domain_16s.fna and placing onto the same archaeal 16S reference alignment and tree, one with barrnap's archaeal 16S profile and one with its bacterial one:
| target | profile | hits | reftreename |
| ------ | ------------------------ | ---- | ----------- |
| arc16s | barrnap arc.hmm 16S_rRNA | 11 | arc16s |
| bac16s | barrnap bac.hmm 16S_rRNA | 12 | arc16s |
The two profiles hit disjoint halves of the 23 search sequences, so the joint outputs demonstrably contain queries neither row yields on its own — verified end to end: the joint classification has 187 rows against 105 + 82 for the two individual ones, over 23 distinct queries against 11 + 12.
Placing bacterial sequences on an archaeal reference tree is deliberate and not biologically meaningful — it puts them on long branches. The point is two different profiles legitimately sharing one reference tree, which is the situation the feature exists for, using reference data already present on this branch. No new data files are added; the sheet only references existing ones.
README.md documents it under the phyloplace section, including the caveat above.
Generated by Claude
nf-core/test-datasetsGitHub
08/31/2026, 5:53 PMGitHub
08/31/2026, 6:58 PMGitHub
08/31/2026, 7:11 PM