MINING-D is a tool (written in Python 3.6.5) for inference of D genes using Rep-Seq data from diverse species. It takes as input a fasta file with consensus CDR3s from an immunosequencing dataset (see paper) and writes the inferred D genes in the output file in the fasta format.
There are two ways of running MINING-D.
-
From command line - To use default MINING-D parameters, use
$ python MINING_D.py -i <input_CDR_file> -o <output_file>To check all available options, use
python MINING-D.py --help.$ python MINING_D.py --help Usage: MINING_D.py [OPTIONS] Options: -i, --input_cdr_file TEXT Input file with consensus CDR3s [required] -o, --output_file TEXT Output file to store inferred D genes [required] -k INTEGER Length of starting k-mers [default: 10] -n, --num_k_mers INTEGER Number of most abundant k-mers to extend [default: 300] -p, --p_val_th FLOAT p-value threshold for the extension procedure [default: 4.5e-36] -t, --n_cores INTEGER Number of cores [default: 10] --bidir_alpha FLOAT Alpha for filtering bidirectional extensions (see paper) [default: 0.5] -g, --cliq_th INTEGER Similarity metric threshold for generating graphs [default: 2] --help Show this message and exit. -
Jupyter Notebook - If you would like to run MINING-D in an interactive environment, there is a Jupyter notebook named "interactive_MINING_D.ipynb" in the repo. Please specify the input file, output file, and other parameters at the top of the notebook like shown in the example.
Dependencies
- Biopython
- Networkx<2.4
- Joblib
- NumPy
- SciPy
- click