Skip to main content

MHC class I binding and epitope prediction based on modeled physicochemical properties of peptides

Project description


MHC class I binding and epitope prediction based on modeled physicochemical properties of peptides.


MHCLovac is a result of my personal interests and some free time. It is not a result of thorough scientific research. That said, MHCLovac is not too bad in terms of predictions it makes and I plan to improve it further if possible.

What's new?

  • version 3.0
    • epitope prediction.
    • ic50 predictions are replaced with binding score. The higher the score the stronger the binding.
    • single prediction algorithm replaced with collection of algorithms, prediction consensus.


MHCLovac is a command line tool used for MHC Class I binding and epitope prediction. It uses modeled physicochemical properites of target peptides to predict binding and epitope scores. The modeling is accomplished using proteinko python package (also my work). MHCLovac is not constrained by the length of target peptide sequence and is capable of making predictions for peptides of any length. Prediction is carried out by a collection of regression and classification algorithms. Algos are trained on data obtained from two sources: dataset used for retraining the IEDB class I binding prediction tools and IEDB database (

Training results and a list of supported MHC alleles is available in training/results folder.

Trained models are benchmarked using ROC-AUC method. Benchmarking method is explained in benchmark folder.



pip install mhclovac

Example usage

mhclovac -f example.fasta -m HLA-B*44:02 -l 11

Example output

 sequence          mhc  peptide_length           sequence_name  binding_score  epitope_score  combined_score
 MEIFIEVFSHF  HLA-B*44:02              11  MEIFIEVFSHF HLA-B44:02       0.523205       0.965484        1.488688
 EIFIEVFSHFL  HLA-B*44:02              11  MEIFIEVFSHF HLA-B44:02       0.087188       0.512132        0.599320
 IFIEVFSHFLL  HLA-B*44:02              11  MEIFIEVFSHF HLA-B44:02       0.039142       0.159362        0.198503
 FIEVFSHFLLQ  HLA-B*44:02              11  MEIFIEVFSHF HLA-B44:02       0.114877       0.264553        0.379430
 IEVFSHFLLQL  HLA-B*44:02              11  MEIFIEVFSHF HLA-B44:02       0.317922       0.964168        1.282090


  1. sequence
  2. sequence_name - Fasta sequence name or name provided by -n argument
  3. peptide_length
  4. mhc - MHC allele
  5. binding_score - Higher score means better binding
  6. epitope_score - Higher score means a better epitope
  7. combined_score - Sum of binding and epitope scores if both are available

Donate to support my work

This work is done on my own budget so if you like you can support me by donating Bitcoin. Any amount donated will be appreciated. Thank you!

BTC: bc1qrg7wku5g35kn0qyay4uwzugfmfqwnvz95g54pj

Project details

Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Files for mhclovac, version 3.1
Filename, size File type Python version Upload date Hashes
Filename, size mhclovac-3.1-py3-none-any.whl (36.6 MB) File type Wheel Python version py3 Upload date Hashes View
Filename, size mhclovac-3.1.tar.gz (36.6 MB) File type Source Python version None Upload date Hashes View

Supported by

AWS AWS Cloud computing Datadog Datadog Monitoring DigiCert DigiCert EV certificate Facebook / Instagram Facebook / Instagram PSF Sponsor Fastly Fastly CDN Google Google Object Storage and Download Analytics Pingdom Pingdom Monitoring Salesforce Salesforce PSF Sponsor Sentry Sentry Error logging StatusPage StatusPage Status page