laboratoriobioinformatica/marvel

MARVEL: Metagenomic Analyses and Retrieval of Viral ELements

marvel's Issues

Prokka results created in the input directory instead of the output directory

Hello!

Thank you for developing MARVEL!

I believe you are running prokka and creating the files in results inside the input_folder and checking (in checkpoints) for the prokka result in outdir. If the input and output directories are different, MARVEL throws the following error.

Traceback (most recent call last):
  File "marvel_bins.py", line 206, in <module>
    bin_dirs = os.listdir(prokka_results)
FileNotFoundError: [Errno 2] No such file or directory: '<output_dir>/prokka'

Shouldn't the prokka result be created in the output directory?

Unable to control output directory

Hello!

Currently, the user cannot control where the output is written. This makes it inconvenient since I (and many others) like to organize using directories. As it is, my input directory (bin directory) gets cluttered with files from MARVEL.

It would be great to have a -o|--outdir argument, where the final output files are placed, with some generic default such as out. Ideally, intermediate files would be written to a temporary directory, or at least to --outdir and deleted upon completion.

If this gets implemented, it is also convenient to create the specified directory if it does not yet exist. E.g.

if not os.path.isdir(args.outdir):
    os.makedirs(args.outdir)

FileNotFoundError: [Errno 2] No such file or directory: 'results/prokka'

I encountered this problem while using it
Traceback (most recent call last):
File "marvel_bins.py", line 205, in
bin_dirs = os.listdir(prokka_results)
FileNotFoundError: [Errno 2] No such file or directory: 'results/prokka'
thanks

Unable to run from other directories - Hard coded models

Hi! I am enjoying using your tool, but I have run into a problem.

I noticed that I cannot run MARVEL from directories other than this repository.

For instance, trying to run from the parent directory:

$ ls
data/ MARVEL/

$ python3 MARVEL/marvel_bins.py -i data/bins_dir/
**Your database and models are not set. Please, run: python download_and_set_models.py

However, when running from this repository, it works:

$ ls
data/ MARVEL/

$ cd MARVEL

$ python3 marvel_bins.py -i ../data/bins_dir    # runs fine

This is because the path to the models is hard-coded in the script, first occuring on line 130:

if not os.path.isfile('models/all_vogs_hmm_profiles_feb2018.hmm'):

Most tools do not require the user to cd into the directory containing the script since it is restrictive. It would be more convenient to provide the models location as a command-line argument:

parser.add_argument(
    '-m',
    '--models',
    default='models/all_vogs_hmm_profiles_feb2018.hmm',
    help='HMM models file')

...

# Verify databases
if not os.path.isfile(args.models):
    parser.error(f'--models {args.models} does not exist. Please specify correct directory, or run python download_and_set_models.py if you have not already')

MARVEL fails at hmmsearch step

I am trying to run MARVEL and it keeps failing at the hmmsearch step.

Letting you know that I was able to fix it. In line 246 of marvel_bins.py, moving the the option -Z 9518 before <hmmfile> and <seqdb> solved the issue.

Thanks!

phage bins were not detected from provided example data

Hi. I'm trying to isolate phage bins in my data using MARVEL, so I installed it and tested with provided example_data.

However, I got this results: we did not find any phage bins in this sample.

I tried to excute with both 8k and 10k bins folders, but I got same issues. (I just follwed the installation and test process in README.)

Here is the result screenshot.

How can I solve this problem? Thank you!