creating conda environment from snakemake rule

Question

[Jordi]

I am trying to activate a conda environment from snakemake to use a different environment for a rule in the workflow. I created a .yml file to specify dependencies.

This is the snakemake rule I defined:

rule Merge_VCFs:
    input:
        f1='file1',
                f2='file2'
    output:
                vcf="output_file"
    conda:
               "bcftools_env.yml"
    shell:
        """
        bcftools merge -m none -F x {input.f1} {input.f2} > {output.vcf}
                """

When I try to run this, however, I get the following error:

Workflow.conda() missing 1 required positional argument: 'conda_env'

I looked for information on this argument in the documentation, but I could not figure out what is missing and where I need to specify it. I ran rules using other .yml previously without encountering this error.

Does anybody have any idea what this is?

These are the run parameters used:

snakemake -s "Snakefile" \
        --cores 10 \
        --resources mem_gb=40 \
        --use-conda \
        --conda-prefix '/home/USER/anaconda3'

This is the content of the .yml file I am trying to use to set up the environment:

name: bcftools_env
channels:
  - conda-forge
  - bioconda
  - defaults
dependencies:
  - _libgcc_mutex=0.1=conda_forge
  - _openmp_mutex=4.5=2_gnu
  - bcftools=1.17=h3cc50cf_1
  - bzip2=1.0.8=h7f98852_4
  - c-ares=1.19.1=hd590300_0
  - ca-certificates=2023.5.7=hbcca054_0
  - gsl=2.7=he838d99_0
  - htslib=1.17=h81da01d_2
  - keyutils=1.6.1=h166bdaf_0
  - krb5=1.20.1=h81ceb04_0
  - libblas=3.9.0=17_linux64_openblas
  - libcblas=3.9.0=17_linux64_openblas
  - libcurl=8.1.2=h409715c_0
  - libdeflate=1.18=h0b41bf4_0
  - libedit=3.1.20191231=he28a2e2_2
  - libev=4.33=h516909a_1
  - libgcc-ng=13.1.0=he5830b7_0
  - libgfortran-ng=13.1.0=h69a702a_0
  - libgfortran5=13.1.0=h15d22d2_0
  - libgomp=13.1.0=he5830b7_0
  - libnghttp2=1.52.0=h61bc06f_0
  - libnsl=2.0.0=h7f98852_0
  - libopenblas=0.3.23=pthreads_h80387f5_0
  - libssh2=1.11.0=h0841786_0
  - libstdcxx-ng=13.1.0=hfd8a6a1_0
  - libzlib=1.2.13=hd590300_5
  - ncurses=6.4=hcb278e6_0
  - openssl=3.1.1=hd590300_1
  - perl=5.32.1=2_h7f98852_perl5
  - xz=5.2.6=h166bdaf_0
  - zlib=1.2.13=hd590300_5
  - zstd=1.5.2=h3eb15da_6
prefix: /home/USER/anaconda3/envs/bcftools_env

[camerond]

Have you added --use-conda flag when running snakemake?

[Jordi]

Yes, I did. I just edited the post accordingly

[camerond]

Should you have envs/bcftools_env.yml next to your conda directive? Or envs at the end of the path after the --conda-prefix flag?

[camerond]

I'm not sure the --conda-prefix is necessary. I ran your code without it, after setting up the conda env, and it ran fine. Had to fix a few typos in the Snakefile though.

[Jordi]

I will try this out soon. It is weird, however, that I ran other rules with similar setups without any problem. Yes, for some reason the spacing in the Snakefile is messed up in my post. The original file is correct, though.

[camerond]

It's not just the spacing, you have a comma missing after f1='file1'. Sometimes when erroneous formatting like this is present snakamake throws ambiguous errors.