jeremycfd
commented
5 years ago TCRdist is using allele-level information to calculate probabilities of generation, among other things. So at the stage of processing a parsed_seqs_file, it's looking for chains that can be explicitly found in the current version of the annotation database, which is found in /tcr-dist/db/alphabeta_db.tsv This is one of the reasons it's generally best to parse/annotate your sequences using the TCRdist pipeline so that you aren't introducing different segment names, etc. Or you can annotate your data using whatever other pipeline you have, but use the TCRdist VDJB database for those annotations. If you can't do that, you can try editing the database file and then using the option --no_probabilities (otherwise I would expect your probabilities to be incorrect.)
Having said all of that... I don't quite understand what you mean by "I currently only have V gene-level information". The example data you posted includes J gene information as well.
Feel free to email me directly if you want to discuss further how to get an unsupported data format into TCRdist. It is possible to do, but it's important to take into consideration that some of the standard analyzes presented in the output are going to be biased as a result.
phbradley
ALGW71
Owuorgpo
Hello,
I currently only have V gene-level information and the first few rows of my "parsed_seqs_file" looks like this table below. My error comes from the "all_genes.py" script not able to find the gene in this file "alphabeta_db.tsv" because it doesn't have the allele information (example: TRAV26-1 not found, when expecting something like TRAV26-1*01). Is there a way to run tcr-dist without this information? Thank you so much for any advice.
My script run is this:
python2 run_basic_analysis.py --organism human --parsed_seqs_file parsed_seqs_file --make_fake_quals --no_probabilities > parsed_seqs_file.out
-Gabrielle