cgMLST schemes?

[slvrshot]

Could you explain a little bit how to download a cgMLST scheme for Neisseria spp.. for use with this tool? I'm interested in doing this for a few samples but the documentation seems relevant only to traditional 7 loci MLST. As a long time StringMLST user thanks for this tool!

[ar0ch]

If you take a look at the indexer docs, they should look very familiar since you're a stringMLST user. If you have a scheme you've built with stringMLST, you can use the same input files for STing. The docs and examples focus around the 7 locus schemes since they tend to be easier to work with and faster to build. Did you have a particular scheme in mind?

[slvrshot]

If you take a look at the indexer docs, they should look very familiar since you're a stringMLST user. If you have a scheme you've built with stringMLST, you can use the same input files for STing. The docs and examples focus around the 7 locus schemes since they tend to be easier to work with and faster to build. Did you have a particular scheme in mind?

Thanks for the rapid response. The Neisseria cgMLST v1.0 cgc400 scheme on pubMLST (Harrison et al.). is what I am trying to work with. One problem I am having is downloading all of the allele sequences. pubMLST seems to limit the number I am able to download at once. The scheme has over 1.6K loci.

https://pubmlst.org/bigsdb?db=pubmlst_neisseria_seqdef&page=batchSequenceQuery

[ar0ch]

It looks like the download links are in: https://pubmlst.org/bigsdb?db=pubmlst_neisseria_seqdef&page=downloadAlleles&locus= format

So let's try something a little brutish to get everything.

First, head over to the export tab and just export the profile table for the STs (saved as profiles.txt): https://pubmlst.org/bigsdb?db=pubmlst_neisseria_seqdef&page=job&id=BIGSdb_043821_1608874811_82059

Extract the header and clean it up:

head -n 1 profiles.txt | tr '\t' '\n' | tail -n +2 | head

Hopefully you see something like: NEIS0001 NEIS0004 NEIS0005 NEIS0006 NEIS0007 NEIS0008 NEIS0009 NEIS0010 NEIS0011 NEIS0012

Download the alleles:

head -n 1 profiles.txt | tr '\t' '\n' | tail -n +2 | xargs -I nmb wget "https://pubmlst.org/bigsdb?db=pubmlst_neisseria_seqdef&page=downloadAlleles&locus=nmb" -O nmb.fasta

That should get you all the files you need. I'm on my phone and haven't tested it though. Good luck!

[ar0ch]

Oh, and merry Christmas 🙂

[slvrshot]

Oh, and merry Christmas 🙂

I figured there was some bash script I could use to pull the allele sequences out! Thanks so much I will test this later today. And a very blessed and Merry Christmas to you too!!!

[slvrshot]

Oh, and merry Christmas 🙂

Hello again! That worked. However I'm having issues with creating the database. First I tried your demo but I get an error message:

db_util.py fetch --query "Neisseria spp." --out_dir my_dbs --build_index
Traceback (most recent call last):
  File "/apps/x86_64/miniconda3/lib/python3.8/urllib/request.py", line 1350, in do_open
    h.request(req.get_method(), req.selector, req.data, headers,
  File "/apps/x86_64/miniconda3/lib/python3.8/http/client.py", line 1240, in request
    self._send_request(method, url, body, headers, encode_chunked)
  File "/apps/x86_64/miniconda3/lib/python3.8/http/client.py", line 1286, in _send_request
    self.endheaders(body, encode_chunked=encode_chunked)
  File "/apps/x86_64/miniconda3/lib/python3.8/http/client.py", line 1235, in endheaders
    self._send_output(message_body, encode_chunked=encode_chunked)
  File "/apps/x86_64/miniconda3/lib/python3.8/http/client.py", line 1006, in _send_output
    self.send(msg)
  File "/apps/x86_64/miniconda3/lib/python3.8/http/client.py", line 946, in send
    self.connect()
  File "/apps/x86_64/miniconda3/lib/python3.8/http/client.py", line 1409, in connect
    self.sock = self._context.wrap_socket(self.sock,
  File "/apps/x86_64/miniconda3/lib/python3.8/ssl.py", line 500, in wrap_socket
    return self.sslsocket_class._create(
  File "/apps/x86_64/miniconda3/lib/python3.8/ssl.py", line 1040, in _create
    self.do_handshake()
  File "/apps/x86_64/miniconda3/lib/python3.8/ssl.py", line 1309, in do_handshake
    self._sslobj.do_handshake()
ssl.SSLCertVerificationError: [SSL: CERTIFICATE_VERIFY_FAILED] certificate verify failed: self signed certificate in certificate chain (_ssl.c:1108)

During handling of the above exception, another exception occurred:

Traceback (most recent call last):
  File "/apps/x86_64/STing/STing/db_util.py", line 272, in <module>
    main()
  File "/apps/x86_64/STing/STing/db_util.py", line 261, in main
    dbFileContent = get_db_file_content(dbFileUrl)
  File "/apps/x86_64/STing/STing/db_util.py", line 74, in get_db_file_content
    response = urllib.request.urlopen(fileUrl)
  File "/apps/x86_64/miniconda3/lib/python3.8/urllib/request.py", line 222, in urlopen
    return opener.open(url, data, timeout)
  File "/apps/x86_64/miniconda3/lib/python3.8/urllib/request.py", line 525, in open
    response = self._open(req, data)
  File "/apps/x86_64/miniconda3/lib/python3.8/urllib/request.py", line 542, in _open
    result = self._call_chain(self.handle_open, protocol, protocol +
  File "/apps/x86_64/miniconda3/lib/python3.8/urllib/request.py", line 502, in _call_chain
    result = func(*args)
  File "/apps/x86_64/miniconda3/lib/python3.8/urllib/request.py", line 1393, in https_open
    return self.do_open(http.client.HTTPSConnection, req,
  File "/apps/x86_64/miniconda3/lib/python3.8/urllib/request.py", line 1353, in do_open
    raise URLError(err)
urllib.error.URLError: <urlopen error [SSL: CERTIFICATE_VERIFY_FAILED] certificate verify failed: self signed certificate in certificate chain (_ssl.c:1108)>

I also tried building my own database file and I get a message stating that I'm missing the loci section. I'm not sure why that is as I followed your example database file format. I've uploaded my file.

Thanks!

test2.txt

[ar0ch]

The first set of errors looks like you're running python on a system with old/outdated (or completely missing) SSL certificates or that's sitting behind an SSL-stripping middleware. Running PYTHONHTTPSVERIFY=0 db_util.py fetch --query "Neisseria spp." --out_dir my_dbs --build_index should bypass that (though, obviously, not recommended as a general usage thing). You might need to install the certificates package (pip install --upgrade certifi and/or conda install -c anaconda certifi)

This config.txt works for me.

[slvrshot]

The first set of errors looks like you're running python on a system with old/outdated (or completely missing) SSL certificates or that's sitting behind an SSL-stripping middleware. Running PYTHONHTTPSVERIFY=0 db_util.py fetch --query "Neisseria spp." --out_dir my_dbs --build_index should bypass that (though, obviously, not recommended as a general usage thing). You might need to install the certificates package (pip install --upgrade certifi and/or conda install -c anaconda certifi)

This config.txt works for me.

Okay strangely...my tab delimited file does not seem to work...but it does work when I delete everything below [loci] in your file and then paste the relative paths to the fasta files in my directories. So I'm almost there. However now I get this error:

Total sequences loaded: 1287967
Loading the profiles file... Done!
terminate called after throwing an instance of 'std::bad_alloc'
  what():  std::bad_alloc
Aborted (core dumped)

If you'd prefer I open a new issue please let me know. Thanks!

[ar0ch]

That's just your machine running out of allocatable memory

[slvrshot]

That's just your machine running out of allocatable memory

Hmmm...okay let me try submitting this as a job.

Thanks!

[slvrshot]

That's just your machine running out of allocatable memory

After submitting this as a job on my HPC...its still creating a database for the 1605 loci in the cgMLST scheme I am working on. So far there are over 500K database files that have been created and it is still not done. I've just left it running, but I was curious if this is normal?

[ar0ch]

That sounds about right for a cg scheme.

[slvrshot]

That sounds about right for a cg scheme.

Happy New Year! Okay I'm almost there....Question I have my database in a folder called "Ng" containing 24K files for 4 alleles....the files look like this:

db.prof_idx.txt.3105   db.txt.923
db.prof_idx.txt.3106   db.txt.924
db.prof_idx.txt.3107   db.txt.925
db.prof_idx.txt.3108   db.txt.926
db.prof_idx.txt.3109   db.txt.927

I just want to verify I am running this correctly, particularly the database prefix portion:

typer -x Ng/db -1 test.sim_1.fastq -2 test.sim_2.fastq -s test -o test.tsv

I should mention that even with these 4 alleles typer seems to be taking a very long time even on the compute cluster. Thanks again!

[slvrshot]

That sounds about right for a cg scheme.

Happy New Year! Okay I'm almost there....Question I have my database in a folder called "Ng" containing 24K files for 4 alleles....the files look like this:

db.prof_idx.txt.3105   db.txt.923
db.prof_idx.txt.3106   db.txt.924
db.prof_idx.txt.3107   db.txt.925
db.prof_idx.txt.3108   db.txt.926
db.prof_idx.txt.3109   db.txt.927

I just want to verify I am running this correctly, particularly the database prefix portion:

typer -x Ng/db -1 test.sim_1.fastq -2 test.sim_2.fastq -s test -o test.tsv

I should mention that even with these 4 alleles typer seems to be taking a very long time even on the compute cluster. Its been stuck on the below for over 2 hrs despite being "Done!" for allele calling after only 10 minutes. Thanks again!

Loading the input index...  Done!
Processing reads...    Done!
Getting the top 2 alleles of each locus...  Done!
Calling alleles based on k-mer frequencies...   Done!

[ar0ch]

Did you build the full profile and did it finish running? A ~1600 loci scheme shouldn't take more than 40-50mins on a 60-70x coverage sequencing run. Admittedly the machine we run things on has a fair amount of memory, but since you could build the database (which takes more memory) you shouldn't have any issues (it'll take ~14GB).

If your reads are a mixture of more than one species though, (1) you shouldn't run them in STing, since it is designed for single sample runs, and (2) the allele selection and refinement step will take a long time

[slvrshot]

I did build the full 1600 loci scheme. Admittedly, I do have 200+ WGS samples. If it takes an hour per sample that's essentially a week and a half of processing. I'll give it a try again.

[slvrshot]

@ar0ch Okay finally ran my samples and it was relatively fast to run all 200. I did notice the profiler didn't predict any STs despite identifying the majority of alleles for each of the 1600+ loci. Happy to share the data. Can I email you? Thanks!

[ar0ch]

Sure -- mail [@] aroonchande.com

[slvrshot]

Sure -- mail [@] aroonchande.com

I messaged you. Did it go through? The first attempt bounced back. Thanks!

[ar0ch]

Got it. I'll try and take a look this evening

[ar0ch]

(I haven't forgeten this issue, just busy with other things)