Importing distance.cmp into R

[UniAlberta]

trafficstars

Hi, I ran sourmash sketch dna ~/*.fastq.gz on my fasta files and then sourmash compare *.sig -o distances.cmp -k 31. Now I have distance.cmp output file and I want to use it in R to plot ordination. I`m not sure how I can import .cmp file into R. Is there any code for that? Thanks

[ctb]

hi @UniAlberta, you'll want to use the --csv output instead - the .cmp file is a numpy binary matrix file that is probably more difficult to read into R!

Here's some example code: https://sourmash.readthedocs.io/en/latest/other-languages.html#r-code-for-working-with-compare-output

[UniAlberta]

Thanks for your reply. I have all my .files and when I try to run sourmash compare *.sig --csv distance.cmp.csv, I get an error of ModuleNotFoundError: No module named 'numpy'. Could you please help to fix it? Here`s the info from my smash (smash) ...@...:~/miniforge3/envs/smash/lib/python3.9/site-packages/sourmash$ ls init.py _lowlevel.py cli commands.py fig.py lca nodegraph.py sbt_storage.py sig utils.py main.py _lowlevel__ffi.py command_compute.py compare.py hll.py logging.py np_utils.py sbtmh.py signature.py version.py pycache _lowlevel__lib.so command_sketch.py exceptions.py index.py minhash.py sbt.py search.py sourmash_args.py

[ctb]

that's weird - what command did you use to install sourmash?

In any case, after activating the conda environment, you should be able to use

pip install numpy

or

conda install numpy

[UniAlberta]

I follow this for installation https://sourmash.readthedocs.io/en/latest/tutorial-install.html. I run conda install numpy and then when I`m in smash environment, I run sourmash compare *.sig --csv distance.cmp.csv and it gives me the same error. ModuleNotFoundError: No module named 'numpy'. Have you updated the tool? Because I ran it last week and it worked

[ctb]

Nope, no updates. And in any case that wouldn't have broken your conda environment!

I'm wondering if maybe your conda environment activation is somehow messed up - try logging in again/starting a new shell, and then activating the smash environment again. conda list should show that numpy is installed, along with sourmash.

[UniAlberta]

I tried pip3 install numpy and it worked. Thanks for your help.

[ctb]

fantastic!

please feel free to ask for help here or elsewhere - more people monitor https://github.com/dib-lab/sourmash/issues so that's a good place to go :)