Skip to content

Troubleshooting Steps

Error 1. Issues with PASTA

Solution

Bioconda's pasta package (>=1.9.0) normally installs run_pasta.py/run_seqtools.py correctly on its own - the rules in workflow/rules/multi_align.smk already call those names directly, so this shouldn't come up in a clean install. If it still does (e.g. run_pasta.py/run_seqtools.py not found, or resolving to the wrong install), build PASTA from source instead. Run the following from the main ROADIES repository directory (after doing cd ROADIES), within the activated Conda environment:

git clone https://github.com/smirarab/pasta.git
git clone https://github.com/smirarab/sate-tools-linux.git
cd pasta
python3 setup.py develop --user

This installs run_pasta.py/run_seqtools.py as scripts under ~/.local/bin, which takes priority on PATH over any conda environment's own copy - including in other conda environments on the same machine later on. If PASTA behaves oddly after switching environments or machines, check for a stray ~/.local/bin/run_pasta.py/run_seqtools.py from a past run of this workaround before assuming something else is wrong.

After doing this change, please re-run the ROADIES pipeline.

Error 2. Environment conflict

Solution

If you encounter the following error message - "ls: relocation error: /lib64/libacl.so.1: symbol getxattr, version ATTR_1.0 not defined in file libattr.so.1 with link time reference", please run the following command to resolve it: 

export LD_LIBRARY_PATH=/usr/lib64/:${LD_LIBRARY_PATH}

Error 3. Mamba not found in the shell

When running the following command: 

$ python ROADIES/run_roadies.py --cores 1
You may encounter this error:

rm: cannot remove '<OUT_DIR>': No such file or directory
Unlocking working directory.
snakemake --cores 1 --config mode=accurate config_path=config/config.yaml num_threads=0 --use-conda --rerun-incomplete
Config file config/config.yaml is extended by additional config specified via the command line.
Building DAG of jobs...
CreateCondaEnvironmentException:
The 'mamba' command is not available in the shell /usr/bin/bash that will be used by Snakemake. You have to ensure that it is in your PATH, e.g., first activating the conda base environment with `conda activate base`.The mamba package manager (https://github.com/mamba-org/mamba) is a fast and robust conda replacement. It is the recommended way of using Snakemake's conda integration. It can be installed with `conda install -n base -c conda-forge mamba`. If you still prefer to use conda, you can enforce that by setting `--conda-frontend conda`.
This means mamba package manager is missing or not available in the environment.

Solution

Install mamba:

conda install -n base -c conda-forge mamba

If you prefer using conda, you can enforce it by adding the --conda-frontend conda argument.

Step 1: In the downloaded ROADIES repository, open the file noconverge.py inside the workflow folder (ROADIES/workflow/noconverge.py).

Step 2: At line 31, add the argument --conda-frontend conda to the cmd command, as shown below:

cmd = [
    "snakemake",
    "--cores",
    str(cores),
    "--config",
    "mode=" + str(mode),
    "config_path=" + str(config_path),
    "num_threads=" + str(num_threads),
    "--use-conda",
    "--rerun-incomplete",
    "--conda-frontend", "conda"
]
Step 3: Rerun the pipeline as follows:

python run_roadies.py --cores 16

Error 4. Conda not recognized

This can happen if conda is not added to your system's PATH.

Solution

To resolve this, please ensure conda is added to the PATH by running the following commands:

export PATH="$HOME/miniconda3/bin:$PATH"
source ~/.bashrc

Error 5. Handling dependencies (glibc)

Solution

Ensure that the glibc version on your system is updated to 2.29 or higher. Update your system libraries if necessary. Otherwise you may encounter this error:

workflow/scripts/lastz_32: /lib64/libm.so.6: version 'GLIBC_2.29' not found

Error 6. PASTA fails with insufficient core count

Pasta fails when the number of cores is insufficient for the number of instances.

The pipeline provides NUM_INSTANCES as a configuration parameter in config.yaml to run multiple instances in parallel. Each instance can also be parallelized using threads. The number of threads per instance is calculated as:

num_threads = number_of_cores / num_instances
If num_instances > number_of_cores, then num_threads will be 0 and the process (e.g., pasta) will fail.

Solution

Ensure that the number of cores is greater than or equal to the number of instances. By default, NUM_INSTANCES is set to 4, so the number of cores (--cores in command line argument) must be at least 4. To run the pipeline with fewer cores, modify the NUM_INSTANCES parameter in the config file:

python run_roadies.py --cores <available_cores> --config_path config/config.yaml

Error 7. MLIPPER fails with undefined symbol: ATL_dGetNB (GPU placement mode)

You may see the following when running --mode placement --gpu:

MLIPPER/MLIPPER: symbol lookup error: /lib/x86_64-linux-gnu/liblapack.so.3: undefined symbol: ATL_dGetNB

This means your system's liblapack.so.3/libblas.so.3 (via Debian/Ubuntu's update-alternatives) currently resolves to an ATLAS build that's missing its own libatlas.so.3 dependency - a broken/incomplete ATLAS package install, unrelated to MLIPPER or ROADIES itself.

Solution

Check which LAPACK/BLAS variant is active:

update-alternatives --display liblapack.so.3-x86_64-linux-gnu
update-alternatives --display libblas.so.3-x86_64-linux-gnu

If a non-ATLAS alternative is listed (commonly under /usr/lib/x86_64-linux-gnu/lapack/ and /usr/lib/x86_64-linux-gnu/blas/), you can point the dynamic linker at it for your ROADIES session without changing the system-wide default:

export LD_LIBRARY_PATH="/usr/lib/x86_64-linux-gnu/lapack:/usr/lib/x86_64-linux-gnu/blas:$LD_LIBRARY_PATH"

Then re-run run_roadies.py. If no non-ATLAS alternative exists on your system, reinstalling libatlas3-base (or your distro's equivalent) should restore the missing libatlas.so.3 dependency.