Files produced are empty post dehosting

[Rohit-Satyam]

Dear @bede

I never had issues with hostile when using single-end ONT data. Today I tried PE fastq files and I see it runs successfully, but the files produced are empty. Any idea what must be causing this?

hostile clean --fastq1 /phd/ultraplex_demux_sample_9_Fwd.fastq.gz --fastq2 /phd/ultraplex_demux_sample_9_Rev.fastq.gz --threads 20 --index human-t2t-hla.rs-viral-202401_ml-phage-202401 -o /phd/dehost --force
22:14:33 INFO: Hostile v2.0.2. Mode: paired short read (Bowtie2)
22:14:33 INFO: Found cached standard index human-t2t-hla.rs-viral-202401_ml-phage-202401
22:14:33 INFO: Cleaning…
22:15:29 INFO: Cleaning complete
[
    {
        "version": "2.0.2",
        "aligner": "bowtie2",
        "index": "human-t2t-hla.rs-viral-202401_ml-phage-202401",
        "options": [],
        "fastq1_in_name": "ultraplex_demux_sample_9_Fwd.fastq.gz",
        "fastq1_in_path": "/phd/ultraplex_demux_sample_9_Fwd.fastq.gz",
        "reads_in": 3469800,
        "reads_out": 1714056,
        "reads_removed": 1755744,
        "reads_removed_proportion": 0.50601,
        "fastq2_in_name": "ultraplex_demux_sample_9_Rev.fastq.gz",
        "fastq2_in_path": "/phd/ultraplex_demux_sample_9_Rev.fastq.gz",
        "fastq1_out_name": "ultraplex_demux_sample_9_Fwd.clean_1.fastq.gz",
        "fastq1_out_path": "/phd/dehost/ultraplex_demux_sample_9_Fwd.clean_1.fastq.gz",
        "fastq2_out_name": "ultraplex_demux_sample_9_Rev.clean_2.fastq.gz",
        "fastq2_out_path": "/phd/dehost/ultraplex_demux_sample_9_Rev.clean_2.fastq.gz"
    }
]

With version 2.0.0 and 1.1.0

mamba create -n hostile -c conda-forge -c bioconda hostile==1.1.0
hostile clean --fastq1 /phd/ultraplex_demux_sample_9_Fwd.fastq.gz --fastq2 /phd/ultraplex_demux_sample_9_Rev.fastq.gz --threads 20 -o /phd/dehost --force
Traceback (most recent call last):
  File "/home/subudhak/miniconda3/envs/hostile/bin/hostile", line 10, in <module>
    sys.exit(main())
             ~~~~^^
  File "/home/subudhak/miniconda3/envs/hostile/lib/python3.13/site-packages/hostile/cli.py", line 170, in main
    defopt.run(
    ~~~~~~~~~~^
        {
        ^
    ...<9 lines>...
        strict_kwonly=False,
        ^^^^^^^^^^^^^^^^^^^^
    )
    ^
TypeError: run() got an unexpected keyword argument 'strict_kwonly'

[bede]

Hi Rohit, would it be possible to share the files with me for testing? My email is in my profile. Thanks

[Rohit-Satyam]

Hi @bede

I just tested if I use single-end mode as follows nd it works. So I am sure now there is an issue with paired-end. I will email you the samples because these need to be confidential and are patient samples.

hostile clean --fastq1 /phd/ultraplex_demux_sample_9_Fwd.fastq.gz --threads 20 -o /phd/dehost

[bede]

Hi Rohit,
As discussed over email, this behaviour relates to the unusual formatting of your paired read headers generated by the tool Ultraplex, leading to your reads not being recognised as paired by Hostile.

Since you have already shared a screenshot of our email conversation containing your read headers, I feel comfortable sharing them again here for clarity:

Your read headers look like this:
@LH00684:44:22TYT2LT3:6:1101:35238:1080_2:N:0:ATTACTCG+AGGCTATArbc:

Hostile expects headers in this (among others) format (note the space in place of the underscore)
@LH00684:44:22TYT2LT3:6:1101:35238:1080 2:N:0:ATTACTCG+AGGCTATArbc:

You can add this space back into your reads prior to processing with Hostile with e.g. sed:

for f in *Fwd.fastq.gz *Rev.fastq.gz; do gunzip -c "$f" | sed 's/_\([12]\):N:0:/ \1:N:0:/' | gzip > "${f%.fastq.gz}.fixed.fastq.gz"; done

If you would rather do this in streaming fashion, you could for example perform streaming interleaving as described here, pipe this through sed as above, and pipe that fixed interleaved fastq stream into hostile in paired mode (untested)

paste <(gunzip -c reads-1.fastq.gz | paste - - - -) <(gunzip -c reads-2.fastq.gz | paste - - - -) | tr '\t' '\n' | sed 's/_\([12]\):N:0:/ \1:N:0:/' | hostile clean --fastq1 - --fastq2 -