install.sh error

[kazumaxneo]

Hi,

I am looking for a tool to handle metagenome analysis including eukaryotic MAG.
I have run the install script in ubuntu18 os and at the end of the installation process, which takes about an hour, I get the following error.

==error message==
EnvironmentLocationNotFound: Not a conda environment:
/home/kazu/miniconda3/envs/mudoger_env/dependencies/conda/envs/otupick_env

cp: target '/home/kazu/miniconda3/envs/mudoger_env/dependencies/conda/envs/otupick_env/bin'
is not a directory
WARNING: The main script cannot be copied into conda environment. Exiting...
==END==

I have miniconda installed in user's home, also installed mamba.
Install script was runned from base enviroment.
Is there anything I can do to solve this error?

[JotaKas]

Hello kazumaxneo,

Thank you for your report.

We will try to replicate the error and get back to you with the fix soon.

[kazumaxneo]

Hello JotaKas

Thank you very much.
I look forward to hearing the results of your test.

[nksjplai]

I am facing the same issue. Any solution will be appreciated. Thanks

[timyerg]

Same issue. Also, many installations above the last errors reported other issues, and the script followed further to other tools. I guess I will get errors downstream during the analyses. Unfortunately, Singularity is not an option for me.

[mdsufz]

Hello everyone,

Unfortunately, the automatic conda installation has shown inconsistencies we could not control. The installation works as intended in some cases, but it may skip some dependency that generates downstream errors in others. That was one of the reasons we implemented the singularity container solution.

Alternatively, you can identify the errors from specific tools, load the relevant conda environment, and manually install the missing dependency.

After running the install.sh, the user will have a conda environment for each integrated tool. If the user wants to customize the use of any tool or fix a dependency issue for a particular situation, it is possible to activate its environments independently.

The environment for each tool is named as "toolname_env" and MuDoGeR attempts to configure the following: bbtools_env brat_env busco_env checkv_env eukcc_env eukrep_env extract_env genemarker_env gtdbtk_env htseq_env khmer_env maker2_env metawrap_env otupick_env prokka_env stampede_clustergenomes_env vcontact2_env vibrant_env virfinder_env virsorter2_env wish_env

Therefore, for instance, for you to activate the BRAT environment, you can do the following:
conda activate path/to/your/miniconda/envs/mudoger_env/dependencies/conda/envs/toolname_env

You should also have the mudoger_env itself available after running the install.sh. Therefore, some relevant paths should already be configured, and you should be able to activate a specific environment by running the following:

$ conda activate mudoger_env
$ config_path="$(which config.sh)"
$ database="${config_path/config/database}"
$ source $config_path
$ source $database
$ conda activate "$MUDOGER_DEPENDENCIES_ENVS_PATH"/toolname_env

Once the relevant environment is activated, you should able to install the missing dependencies using conda install.

For instance, from my experience, one problematic environment is the metawrap. To fix the metawrap env, you should activate the
metawrap_env and follow the instructions from its source tool. In this case MetaWrap.

Finally, regarding the otupick_env initial issue, you should be able to do something similar. However, you should have a file containing the dependencies from the otupick_env after you have cloned the mudoger repository. The file should be located at MuDoGeR/installation/dependencies/otupick_dependencies/environment.yml.
Therefore, for the otupick_env you can run the following:

$ conda activate mudoger_env
$ config_path="$(which config.sh)"
$ database="${config_path/config/database}"
$ source $config_path
$ source $database
$ conda env create --prefix $MUDOGER_DEPENDENCIES_ENVS_PATH/otupick_env --file $MUDOGER_DEPENDENCIES_PATH/otupick_dependencies/environment.yml
$ mamba install -y --prefix $MUDOGER_DEPENDENCIES_ENVS_PATH/otupick_env -c anaconda gawk
$ chmod +x $MUDOGER_DEPENDENCIES_PATH/otupick_dependencies/*
$ cp -r $MUDOGER_DEPENDENCIES_PATH/otupick_dependencies/* $MUDOGER_DEPENDENCIES_ENVS_PATH/otupick_env/bin