[khmer] Extracting kmer sequences from khmer output

[Dooley, Damion]

Hi folks,

I think my previous install a few months ago suffered from one point - we'd used the suggested virtualenv pip install khmer to get a standard install, but discovered that part or all of the /sandbox/ folder didn't show up in that.  So I did a separate direct github install which included sandbox, but which didn't compile c code successfully.  I wasn't sure if that was important or not (i.e. precompiled c binaries already in place?)

Long and short of it: should I be doing a python 2.7 virtualenv, loading up the latest github master, and resolving any "make" issues (i didn't see a ./configure file so is there other necessary config there?); or should I install via standard pip approach and copy the sandbox folder in separately?

Thanks for the info,

Damion

Hsiao lab, BC Public Health Microbiology & Reference Laboratory, BC Centre for Disease Control
655 West 12th Avenue, Vancouver, British Columbia, V5Z 4R4 Canada
________________________________________
From: C. Titus Brown [ctbrown at ucdavis.edu]
Sent: Tuesday, September 22, 2015 11:31 AM
To: Miller, Ruth
Cc: titus at idyll.org; khmer at lists.idyll.org; Dooley, Damion
Subject: Re: [khmer] Extracting kmer sequences from khmer output

Hi Ruth,

this indicates a version mismatch - the script is from khmer 2.0, but
the khmer library that it is importing is from your previous intsallation.

Try doing

python -c "import khmer; print khmer"

to see where the default khmer import is.  You might need to set your
PYTHONPATH variable to point at the directory above 'khmer/__init__.py'
for 2.0.

best,
--titus

On Tue, Sep 22, 2015 at 11:27:15AM -0700, Miller, Ruth wrote:
> Hi,
>
> Thanks for getting back to me. I am glad that this functionality is
> available.
> I downloaded count-kmers.py, but I am having some problems, because it
> seems to require a module that I don?t have, namely load_countgraph,
> because I get the error:
>
> hashtable from kmers.g
> Traceback (most recent call last):
>   File "/usr/local/packages/khmer/khmer/sandbox/count-kmers.py", line 105,
> in <module>
>     main()
>   File "/usr/local/packages/khmer/khmer/sandbox/count-kmers.py", line 79,
> in main
>     countgraph = khmer.load_countgraph(
> AttributeError: 'module' object has no attribute ?load_countgraph'
>
>
> I have downloaded the latest version of khmer, so I don?t think that is
> the problem.  And I can?t see a script called load_countgraph in the
> sandbox downloads. Is load_countgraph available somewhere?
>
> Thanks,
>
> Ruth
>
> On 2015-09-19, 6:08 AM, "C. Titus Brown" <ctbrown at ucdavis.edu> wrote:
>
> >On Fri, Sep 18, 2015 at 03:25:39PM -0700, Miller, Ruth wrote:
> >> Hi,
> >>
> >> I am hoping to get a list of the sequence of each kmer identified from
> >>my dataset and it?s abundance. This would allow me to compare the
> >>abundance of different kmers in my sample set, to see whether certain
> >>samples cluster together based on the abundance of kmers present.
> >>
> >> Is there a way to do this in khmer?
> >>
> >> Thanks,
> >>
> >> Ruth
> >
> >Hi Ruth,
> >
> >yep! In khmer 2.0 we've added the 'count-kmers.py' and
> >'count-kmers-single.py'
> >scripts in the sandbox/ directory; you can either do
> >
> >     load-into-counting.py kmers.g file1.fq file2.fq file3.fq ...
> >     sandbox/count-kmers.py kmers.g file1.fq
> >
> >if you have multiple files, or
> >
> >     sandbox/count-kmers-single.py file1.fq
> >
> >The graph size parameters (-M, or -x / -N) still apply for setting the
> >size of the kmer countgraph database; see
> >
> >     http://khmer.readthedocs.org/en/v2.0/user/choosing-table-sizes.html
> >
> >for more info.
> >
> >Now, the other problem is that count-kmers.py/count-kmers-single.py aren't
> >installed with 'pip install khmer' because they're not yet part of the
> >supported scripts.  So you'll need to grab them from the URLs below,
> >
> >https://github.com/dib-lab/khmer/blob/master/sandbox/count-kmers.py
> >https://github.com/dib-lab/khmer/blob/master/sandbox/count-kmers-single.py
> >
> >(click 'Raw' to download the text file) *or* check out the git repository
> >and call them from sandbox.
> >
> >Let me know if you have any problems --
> >
> >best,
> >--titus
> >--
> >C. Titus Brown, ctbrown at ucdavis.edu
>

--
C. Titus Brown, ctbrown at ucdavis.edu