Running with singularity

[jaesvi]

Hi, I am trying to run ClairS with singularity, but it keeps looking for a model and an executable in my ${CONDA_PREFIX} . First I had to tweak a bit the command from the README, since -v is docker syntax, not singularity, the SIF file goes after the options and you are specifying it twice. This is my first command:

singularity exec \
-B ${TUM_INPUT}:${TUM_INPUT},${GL_INPUT}:${GL_INPUT},${REF_DIR}:${REF_DIR},${OUT}:${OUT} \ 
clairs_latest.sif \
/opt/bin/run_clairs \
--tumor_bam_fn ${TUM_INPUT}/test.sorted.bam \
--normal_bam_fn ${GL_INPUT}/test.gl.sorted.bam \
--ref_fn ${REF_DIR}/Homo_sapiens_assembly38.fasta \
--threads 32 \
--platform ont_r9 \
--output_dir ${OUT}

After which I get an error: [ERROR] file ${CONDA_PREFIX}/bin/clairs_models/ont_r9 not found I then unpacked the models in the folder where ClairS is looking for them but I still get an error: [ERROR] Cannot find clair3 main entry in ${CONDA_PREFIX}/bin So is the docker/singularity image self contained? I also tried installing through conda but the environment does not solve with incompatible specifications. I am not an advanced user of singularity, so might be doing something wrong.

[zhengzhenxian]

Hi,

Seem the singularity sandbox cannot find the correct conda environment, We added an option(--conda_prefix /opt/conda/envs/clairs) in singularity workflow. Please try again and kindly let me know if the issue persists, thanks.

[jaesvi]

It did run like this. Thank you for your quick help.