Understanding the model predictions- not CDR3 rank not influenced by HLA or peptide

tianshilu / pMTnet

Deep Learning the T Cell Receptor Binding Specificity of Neoantigen

GNU General Public License v2.0

76 stars 20 forks source link

Understanding the model predictions- not CDR3 rank not influenced by HLA or peptide #19

Open ieuangw opened 8 months ago

ieuangw commented 8 months ago

Hi,

I am trying to run pMTnet and notice that each CDR3 always has the same Rank, regardless of which HLA or Peptide it is compared to? These are Neoantigen Peptides so will never have been seen by the model. The run code is below:

python pMTnet.py -input ~/pMTnet_input.csv -library library -output test/output/CAR108predict -output_log test/output/CAR108predict/output.og

I downloaded pMTnet using git clone. Is there a bug or something specific I am not understanding?

Thanks for your help

wtwt5237 commented 5 months ago

Hi @ieuangw

Could you please try our new pMTnet vesion and see if you have the same problem?

https://dbai.biohpc.swmed.edu/pmtnet/analysis-omni.php

Thanks!

Tao