Python script to search TCRs with specific CDR3 sequences in the iReceptor Gateway (https://gateway.ireceptor.org)
USAGE
python3 search_ireceptor.py --cdr3_file cdr3_file.csv --out_folder out_results_folder
--cdr3_file: file containing the CDR3 sequences (see ./test.csv for the format)
--out_folder: output folder where the results will be stored. It contains a) a pickle dictionary when all results are stored together b) a csv file for each TCR containig information about where the TCR was observed (e.g. Pubmed number, patient status etc. etc.)
example usage
Install the requirements.txt file (a virtual environment is recommended), and then run:
python3 search_ireceptor.py --cdr3_file test.csv --out_folder results_test
The results will be stored in ./results_test folder. You'll find a .csv file for each TCR that had at least one hit in iReceptor. For example './results/data_TRA_CAVSKGGGNKLTF.csv' contains the results of the analysis for the alpha chain sequence CAVSKGGGNKLTF (i.e. in which studies TCR sequences with CAVSKGGGNKLTF as CDR3 have been observed).
I usually filter out TCRs that have exactly the the same V,J genes as those in cdr3_file.csv, and I keep ONLY ONE sequence for repertore_id (I'm interested in clonotype, i.e. whether a specific TCR sequence has been observed in other patients)
If available, also dataset for paired TCRs will be generated, but there is relatively few single cell data for TCRs.
N.B. --same_VJ option There is also an option to select only TCR with the same V,J genes as those in cdr3_file.csv [--same_VJ] direcly when running the script ( python3 search_ireceptor.py --cdr3_file test.csv --out_folder results_test --same_VJ) but there may be some format issues: iReceptor uses IMGT nomenclature so genes format like TRAV-8*04 are OK, but TRAV08*04 or TRAV8-2/4-8 are not. Please check the format of the V,J genes in your dataset if using this option.