Invalid BUSCO DATASET

[CaseysC]

I'm running into problems with the Busco database.
It downloaded the list of current datasets in Busco_downloads/information which are at odb12.
However, it seems that check.inputs.py line 163 is hard coded to odb10.

Whatever I do I get:
Invalid BUSCO DATASET: ascomycota_odb12. Run busco --list-datasets to get a full list available datasets

busco --list-dataset list odb12 but they are not be accepted as input.

Would modifying check.inputs.py line 163 fix this bug?

Thank you for building this pipeline!

[AlisonMacFadyen]

Hi @CaseysC,

Not sure if you are still having issues but I thought I would share what worked for me. I have been trying to get FunGAP to run and also encountered issues associated with BUSCO and the available datasets.

Yes, changing line 163 of check_inputs.py would work. However, as far as I can tell, you won't be able to use a dataset flagged "odb12" with the FunGAP installed BUSCO, as it is version 5.1.2 and is not compatible with the odb12 datasets. At least that is what I find, as I obtain this error:

WARNING:        There is a conflict in your config. You specified a dataset from odb12 while simultaneously requesting the datasets_version parameter be odb10. Proceeding with the lineage dataset as specified from odb12
ERROR:  BUSCO v5 only works with datasets from OrthoDB v10 (with the suffix '_odb10'). For a full list of available datasets, enter 'busco --list-datasets'. You can also run BUSCO using --auto-lineage, to allow BUSCO to automatically select the best dataset for your input data.
ERROR:  BUSCO analysis failed !
ERROR:  Check the logs, read the user guide (https://busco.ezlab.org/busco_userguide.html), and check the BUSCO issue board on https://gitlab.com/ezlab/busco/issues

My fix for this issue was to change the check_inputs.py script so that it doesn't attempt to run busco --list-datasets, but provide a dataset file that can be read to see if your chosen dataset is within the list.

def check_busco_dataset(busco_dataset):
    '''Check BUSCO dataset'''
    busco_dbs = []
    busco_dataset_found = False

    with open("/workspace/FunGAP/busco_datasets.txt", 'r') as busco_list:
        for line in busco_list:
            busco_dbs.append(re.findall(r'\S+_odb10|\S+_odb12', line))
    
    print("Loaded 'busco_datasets.txt'.")
    print("Now comparing datasets to user provided BUSCO dataset.")
    
    for dataset in busco_dbs:
        if busco_dataset in dataset:
            busco_dataset_found = True
            
    if busco_dataset_found:
        print('BUSCO_DATASET is ok...')
    else:
        sys.exit(
                 '[ERROR] Invalid BUSCO DATASET: {}. Run busco --list-datasets to '
                 'get a full list available datasets'.format(busco_dataset)
            )

I obtained that dataset list by running busco --list-datasets and then copying that file into the Docker container when building it. Note you need to modify the list so that all the odb12 are odb10. You also need to copy the modified check_inputs.py.

I've found this fixes the issue of not finding the required dataset and making sure you have a compatible dataset list. I don't know if it would work when it comes to downloading the dataset with BUSCO, as I am running the pipeline within an environment without internet access. However, I have noticed that if the BUSCO step fails, FunGAP still "completes".

My forked repo is here in case that is of use, it has a few more changes which are all located in the docker directory:

https://github.com/AlisonMacFadyen/FunGAP_offline/tree/master

All the best,

Alison

[mbnmbn00]

Hi,
Sorry for late response. I actually was working on this and turned out there are a lot to update.
I left academia, so cannot actively work on this project. But I hope I can work on this project again with better design.

[chongjing]

I got same issue, here is how I solved this - just in case someone needed it:

1. I disable the busco datasets checking in check_inputs.py;
2. download BUSCO datasets before building singularity container, then put it to BUSCO’s expected location during building singularity container.

Here is my singularity def file:

Bootstrap: docker
From: taniguti/fungap-base:v1.1.1

%environment
        export FUNGAP_DIR=/workspace/FunGAP
        export AUGUSTUS_CONFIG_PATH=/opt/conda/config
        export AUGUSTUS_SCRIPTS_PATH=/opt/conda/bin
        export PATH=/workspace/FunGAP/:$PATH

%files
        gmes_linux_64_4.tar.gz /workspace/FunGAP/external/
        fungap.conf /workspace/FunGAP/
        fungi_odb10.2024-01-08.tar.gz /workspace/FunGAP/external/
        ascomycota_odb10.2024-01-08.tar.gz /workspace/FunGAP/external/
        basidiomycota_odb10.2024-01-08.tar.gz /workspace/FunGAP/external/
        stramenopiles_odb10.2024-01-08.tar.gz /workspace/FunGAP/external/
        nematoda_odb10.2024-01-08.tar.gz /workspace/FunGAP/external/

%post
        cd /workspace/FunGAP/external/
        tar -zxvf gmes_linux_64_4.tar.gz

        cd gmes_linux_64_4/
        cp other/reformat_fasta.pl .
        perl change_path_in_perl_scripts.pl "/usr/bin/env perl"

        chmod -R o+rwX /opt/conda/config


    # --- Install BUSCO odb10 datasets ---
        FUNGAP_BUSCO_DIR="/workspace/FunGAP/busco_downloads"
        BUSCO_DATA_DIR="/opt/conda/share/busco"
        mkdir -p "$FUNGAP_BUSCO_DIR"

    # Extract each dataset into the BUSCO data directory
        tar -xzf /workspace/FunGAP/external/fungi_odb10.2024-01-08.tar.gz -C "$FUNGAP_BUSCO_DIR"
        tar -xzf /workspace/FunGAP/external/ascomycota_odb10.2024-01-08.tar.gz -C "$FUNGAP_BUSCO_DIR"
        tar -xzf /workspace/FunGAP/external/basidiomycota_odb10.2024-01-08.tar.gz -C "$FUNGAP_BUSCO_DIR"
        tar -xzf /workspace/FunGAP/external/stramenopiles_odb10.2024-01-08.tar.gz -C "$FUNGAP_BUSCO_DIR"
        tar -xzf /workspace/FunGAP/external/nematoda_odb10.2024-01-08.tar.gz -C "$FUNGAP_BUSCO_DIR"

        BUSCO_SYSTEM_DIR="/opt/conda/share/busco"
        mkdir -p "$BUSCO_SYSTEM_DIR"
        cp -r "$FUNGAP_BUSCO_DIR"/* "$BUSCO_SYSTEM_DIR"

    # Clean up
        rm -f /workspace/FunGAP/external/*odb10*.tar.gz

            # Ensure permissions
        chmod -R a+rX "$FUNGAP_BUSCO_DIR"

        # Patch check_inputs.py to skip BUSCO dataset validation via CLI
        sed -i 's/if busco_dataset not in set(busco_dbs):/if False:  # patched: skip BUSCO dataset validation/' /workspace/FunGAP/check_inputs.py