DRAMP download error

[jasmezz]

EDIT: PLEASE SEE #423 (comment) FOR LATEST STATUS

Description of the bug

The DRAMP database issue reported by J.C. Ruitenberg (https://nfcore.slack.com/archives/C02K5GX2W93/p1728552583863299) seems to affect the DRAMP download and processing permanently since it was updated on September 8 (see: http://dramp.cpu-bioinfor.org/static/update.php) – 3 days after funcscan 2.0.0 release.

I reported the problem already to the DRAMP maintainers; if this is not going to be solved soon, the DRAMP download script bin/ampcombi_download.py should be updated on our side to not crash if non-ascii-characters are found in the database sheet during parsing.

Example error here: https://github.com/nf-core/funcscan/actions/runs/11293369456/job/31411353829?pr=422

Command used and terminal output

No response

Relevant files

No response

System information

No response

[jasmezz]

Still no reply from DRAMP maintainers (neither to my e-mail nor via uploading a fixed DRAMP database). I think we should go ahead and adapt the DRAMP download script, @Darcy220606 👾
If you check their homepage, they have it under CC-BY license. Meaning we could upload it on our side (Zenodo?) to track versions and provide a working DRAMP version. Either:

• the June 22 version (which worked fine) or
• the September 8 version without the faulty AMP sequence (DRAMP31926, sequence "KRFKKFFKKVKKSVKKRLKK工FKKPMVIGVTIPF")
  (See http://dramp.cpu-bioinfor.org/static/update.php)

[jasmezz]

Update: DRAMP maintainers fixed the reported error (removed invalid 工 character from sequence DRAMP31926) and updated the database online. However, it turned out that other amino acid sequences contain invalid characters as well (see screenshot). Waiting for updates on that. In any case, @Darcy220606 will adapt the DRAMP download script to filter such AMPs out to not have the pipeline crash when parsing the database.

[jfy133]

Where on early are these characters coming from?!

It's sort of making me question the reliability of the database...

[swelbo]

Hello - has there been any progress with this?

I'm getting this error when running:

nextflow run nf-core/funcscan -profile test,docker --outdir ~/path/to/test

I assume this is the same issues.

#######

Command error:
Traceback (most recent call last):
File "/home/harry/.nextflow/assets/nf-core/funcscan/bin/ampcombi_download.py", line 78, in
download_DRAMP("amp_ref_database")
File "/home/harry/.nextflow/assets/nf-core/funcscan/bin/ampcombi_download.py", line 49, in download_DRAMP
for record in seq_record:
File "/usr/local/lib/python3.11/site-packages/Bio/SeqIO/Interfaces.py", line 72, in next
return next(self.records)
^^^^^^^^^^^^^^^^^^
File "/usr/local/lib/python3.11/site-packages/Bio/SeqIO/FastaIO.py", line 246, in iterate
Seq(sequence), id=first_word, name=first_word, description=title
^^^^^^^^^^^^^
File "/usr/local/lib/python3.11/site-packages/Bio/Seq.py", line 2034, in init
self._data = bytes(data, encoding="ASCII")
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
UnicodeEncodeError: 'ascii' codec can't encode character '\u03a6' in position 1: ordinal not in range(128)

[Darcy220606]

Hi @swelbo Indeed, we are currently working on a fix for ampcombi, so it should be fixed in the coming two weeks. :)

[davidlyttle]

Just bumping the thread, since I'm experiencing the exact same issue -any updates?

[jasmezz]

As discussed on Slack, a fix incl. the new version of ampcombi will be out soon (aim next week). In the meantime, people can use a working DRAMP version from earlier this year; it is already processed by ampcombi, so just extract this archive and supply the resulting directory (or rather the subdirectory within it) to the pipeline with --amp_ampcombi_db <path_to_db_dir>.