aph(3'')-1b not shown

[Alan]

Hi, I am running the latest ResFinder both locally and remotely. But found that there are discrepancy in the results.

From the web ResFinder, it detect aph(3'')-1b but which is not reported in the txt files generated locally. I look into the xml file generated locally, found it did return blast hits for this gene. Apart from aph(3'')-1b, the remaining resistance genes are reported normally in the txt files. I have attached the xml file for your reference.

I run the resfinder using python3 and cgecore was installed. Did I miss anything?

[[e-mail]]

I observed a similar (maybe the same) issue running resfinder.py locally. A considerable number of hits reported in the temporary output files (tmp/.xml) does not show up in the final results (text files). For example, I do not find any hits for sul in file "results_tab.txt" although many hits are registered in "out_sulphonamide.xml". I can upload the respective files, if necessary.

It might be helpful to say that my sequences are quite short (about 150 bp only).

I thought it might be related to the filters, so I tried with several combinations of "threshold" and "min_cov" but without success. Nevertheless, looking through the code (as a non-python programmer), I found in line 158 of "cgecore/blaster/blaster.py" the statement

                 if (cut_off and perc_coverage > 20) or cut_off == False:

which looks a bit suspicious to me because of the hard-coded 20. Is this intended or should a comparison be made to "min_cov" as, e.g. in line 224 ? Anyway, it seems that this is not the cause of the issue.

[RolfKaas]

Dear Alan and David,

Thanks for reporting these bugs. We have someone looking in to this.

I have rewritten ResFinder and we will soon publish this newer version (4.0.0). I would be very interested if this bug has carried over to version 4. Could I get you to test this? Either by the web tool (https://cge.cbs.dtu.dk/services/ResFinder-4.0/) or the using the branch called "4.0"?

Thanks again for helping improving ResFinder.

Best regards, Rolf Sommer Kaas

[[e-mail]]

Dear Rolf,

yes, I think I can run a test with version 4 (from bitbucket) when it is released.

David

[Alan]

Dear Dr. Kaas,

I'm happy to test it but I am not sure how to clone the version 4.0. I have tried to download the resfinder.py in the branch 4.0 and run it locally. The concerned gene didn't show up in the txt files.

Best regards, Alan

[RolfKaas]

Dear Alan,

On the ResFinder main bitbucket page, just under the title "Resfinder" there's a grey box that says "master", if you click that you can choose "4.0" and the instructions for installing version 4.0 will be presented below.

If possible can you send me the fasta file that is causing this issue?

[Alan]

I have successfully installed resfinder v4.0 Error "Could not locate ResFinder database index path: ..../resfinder_db/kma_indexing/" occur even I have defined the blast path using -b.

I'm using the db of resfinder 3.1.

[RolfKaas]

You may need to specify the path to the ResFinder database using the flag "-db_res /path/to/resfinder_db". Does that work?

[RolfKaas]

Another option can also be to create a link/shortcut named "db_pointfinder" in your resfinder directory that points to the location of the database.

[Alan]

I have specified the db using either -db_res or --db_path_res, both return the same error.

This is my command: python3 run_resfinder.py -o test -s "Escherichia coli" --acquired -db_res '/home/resfinder/resfinder_db/' -fa contigs.fasta -b '/home/blast/blastn'

[RolfKaas]

Have you done the following? cd /home/resfinder/resfinder_db/ python3 INSTALL.py

[Alan]

No, I didn't. Because kma is not mandatory with version 3.1. The program run without error after I have index the db using kma. But the aph-3''-1b also not shown in the table. I am now sending the result files to your email. And I have sent the concerned contigs to your email on 22 March. Thank you so much.

[RolfKaas]

Ah, yes, thats a mistake. I've fixed it now, so that you do not need to install KMA if you are just blasting.

[RolfKaas]

ResFinder does not allow overlap between hits in the newest versions, not even very small overlaps.

The web service still allows small overlaps, until it is updated.

The missing hit reported in this issue is overlapping by 1 nucleotide by the aph(6) hit.

It is currently being discussed if small overlaps should be allowed again. No matter what the decision is, we will implement it as an option in version 4, at least for the local installation.

[Alan]

Thank you for the explanation and update!

[RolfKaas]

ResFinder version 3.1.1 and version 4.0.0 now allows 30 bp overlap by default, solving the issue with the aph genes.

An option has also been implemented to change the size of the allowed overlap (the --acq_overlap flag).

[Christina Ahlstrom]

Thanks for updating to allow overlaps. However, the updated python script is not working for me.

[cahlstrom@yeti-login20 ~] python3 /home/cahlstrom/resfinder/resfinder.py -i /home/cahlstrom/results/AK/A1_063/A1_063assembly.fasta -o /home/cahlstrom/results/AK/A1_063/Resfinder -p /home/cahlstrom/resfinder/resfinder_db -b /home/cahlstrom/ncbi-blast-2.9.0+/bin/blastn -t 0.9000 -l 0.60 --acq_overlap 30

Traceback (most recent call last):

File "/home/cahlstrom/resfinder/resfinder.py", line 563, in <module>

allowed_overlap=args.acq_overlap)

File "/home/cahlstrom/resfinder/resfinder.py", line 75, in blast

allowed_overlap=allowed_overlap)

TypeError: __init__() got an unexpected keyword argument 'allowed_overlap'

Any suggestions?
Thanks!
Christina

[RolfKaas]

Dear Christina,

You may need to update the python module named cgecore, it needs to be version 1.5.0.

You update the module by executing:

pip3 install --upgrade cgecore