These subdirectories contain format parsing code for different TCR, peptide, MHC triad input sources

All formatters output into the uniform CSV format:

job_name,cognate,peptide,mhc_class,mhc_1_chain,mhc_1_species,mhc_1_name,mhc_1_seq,mhc_2_chain,mhc_2_species,mhc_2_name,mhc_2_seq,tcr_1_chain,tcr_1_species,tcr_1_seq,tcr_2_chain,tcr_2_species,tcr_2_seq

job_name: A unique name for the row for downstream use

cognate: "true" or "false" string indicating if the triad binds or not

peptide: The peptide's amino acid sequence

mhc_class: "I" or "II"

mhc_<1/2>_chain: "alpha"/"beta"/"heavy"/"light"

tcr_<1/2>_chain: "alpha"/"beta"

<type>_<1/2>_species: Species of origin for the chain ("human"/"mouse")

<type>_<1/2>_seq: Amino acid sequence for the triad chain

Additionally, for pMHC-only tasks, just exclude the tcr-related columns (for example, for use with af3-nf pMHC-only prediction pipelines)

Installation

Conda env

First, create a conda environment:

conda create -n tcr_format_parsers
conda env update -n tcr_format_parsers --file devtools/conda-envs/test.yaml

Then, clone this repo and install:

git clone https://github.com/ljwoods2/tcr_format_parsers
cd tcr_format_parsers
pip install -e .

Finally, download the necessary IMGT data for stitchr to work

# other species must be downloaded if needed
stitchrdl -s human 

Pip only

Install ANARCI.

Then, clone this repo and install:

git clone https://github.com/ljwoods2/tcr_format_parsers
cd tcr_format_parsers
pip install -e .

Finally, download the necessary IMGT data for stitchr to work

# other species must be downloaded if needed
stitchrdl -s human