<#2277 phyloplace: hmmdomains fixtures and gzipped...
# test-data
g
#2277 phyloplace: hmmdomains fixtures and gzipped copies of the PF14720 fixtures Pull request opened by erikrikarddaniel Description Two unrelated additions to the
phyloplace
branch, bundled into one PR so they take one review and one merge. 1.
testdata/hmmdomains/
— Pfam profiles and a two-domain protein
For nf-core/phyloplace's
hmmer/hmmdomains
module test: the
Cytochrom_C
(PF00034.27) and
Cytochrome_CBB3
(PF13442.12) Pfam profiles, and UniProt A0R050, the QcrCC cytochrome c subunit of Mycolicibacterium smegmatis MC2 155. A0R050 carries two cytochrome c domains and both profiles match both of them, scoring in opposite directions — PF00034 wins the first domain, PF13442 the second. That is the case the module exists for: picking the best profile per stretch of a sequence rather than per sequence. Why these aren't fetched at test time The test previously downloaded both profiles from InterPro's
?annotation=hmm
endpoint during the run. That endpoint is intermittently returning HTTP 404 with
{"Error":1040}
— the same URL failed twice and succeeded once inside twenty seconds, for both accessions and for both the trailing-slash and no-slash forms. With two profiles to fetch, a test run succeeded perhaps one time in nine. Moving them here also matches the rule that pipeline repos carry no test data of their own. Contents •
testdata/hmmdomains/PF00034.hmm
,
testdata/hmmdomains/PF13442.hmm
— from InterPro, decompressed, otherwise unaltered. Stored uncompressed so
hmmsearch
reads them directly without depending on a
.gz
suffix surviving however the file is staged. •
testdata/hmmdomains/A0R050.faa
— from UniProt, unaltered. 2.
testdata/gzipped/
— gzipped copies of the PF14720 fixtures
For nf-core/phyloplace #76, which makes the pipeline accept compressed input for every file parameter. Testing that needs a run driven entirely from compressed files, so these are gzipped copies of fixtures the existing phyloplace test profiles already use. Contents •
testdata/gzipped/PF14720_3_sequences.faa.gz
,
PF14720_seed.alnfaa.gz
,
PF14720_seed.faa.gz
,
PF14720_seed.ft.LGCAT.newick.gz
,
PF14720.hmm.gz
— byte-for-byte the corresponding files one directory up, compressed with
gzip -n
so the archives carry no timestamp and stay reproducible. •
testdata/gzipped/PF14720_seed.taxonomy.tsv.gz
— the taxonomy strings embedded in `PF14720_seed_embedded_taxonomy.alnfaa`'s headers, written out as a two-column
id<TAB>taxonomy
table. This is the one input with no uncompressed standalone equivalent on the branch. The content deliberately duplicates the uncompressed fixtures rather than introducing new sequences: what is under test is the handling of the compression, so any difference in the resulting placement would be noise. Documentation Both sets are documented in
README.md
, under the existing
## Datasets for nf-core/phyloplace
section. 🤖 Generated with Claude Code https://claude.ai/code/session_01Dpvh3T7aVgU5XyjqcYQLQD nf-core/test-datasets
Review on #2277 phyloplace: hmmdomains fixtures and gzipped copies of the PF14720 fixtures Review approved by nschan on erikrikarddaniel's pull request Some typos and incorrect taxonomy? nf-core/test-datasets