PELC (Python Eplet Load Calculator)

Overview

PELC is a Python package designed to calculate efficiently the HLA Eplet Load (based on the EpRegistry database) between donors and recipients by loading in a pandas.DataFrame in eplet_comparison.compute_epletic_load the recipients' and donors' typings. See minimal reproducible example for more details.

Getting started

Install from PyPI (recommended)

To use pelc, run pip install pelc in your terminal.

Usage

a. Comparing two alleles

Here is a minimal example of how to use pelc to compare two alleles:

from pelc.simple_comparison import simple_comparison

simple_comparison(
    "A*68:01",
    "A*68:02",
    "output",  # file will be saved as output.csv in the current directory
    verifiedonly=False,  # if True, only verified eplets will be considered, otherwise all eplets will be considered
    interlocus2=True  # doesn't matter for class I alleles
)

In the output.csv file created in the current directory, you will find two rows: "In A*68:02 but not in A*68:01" and "In A*68:01 but not in A*68:02"

b. Batch mode

Here is a minimal example with the file Template.xlsx (click to download):

import pandas as pd

from pelc import batch_eplet_comp, batch_eplet_comp_aux, output_type

if __name__ == "__main__":
    input_path: str = "Template.xlsx"

    output_path: str = "MyOutput"
    input_df: pd.DataFrame = pd.read_excel(
        input_path,
        sheet_name="My Sheet",
        index_col="Index"
    )

    donordf: pd.DataFrame
    recipientdf: pd.DataFrame
    donordf, recipientdf = batch_eplet_comp_aux.split_dataframe(input_df)

    batch_eplet_comp.compute_epletic_load(
        donordf,
        recipientdf,
        output_path,
        output_type.OutputType.DETAILS_AND_COUNT,
        class_i=True,  # Compute class I eplets comparison?
        class_ii=True,  # Compute class II eplets comparison?
        verifiedonly=False,  # How should the epletic charge be computed? Verified eplets only? Or all eplets?
        exclude=None,  # list of indices to exclude
        interlocus2=True  # whether or not to take into account interlocus eplets for HLA of class II
    )

Advanced usage:

a. Not taking into account all loci (if they are not typed for example)

If one wants to determine the eplet mismatches between a donor and a recipient but without taking into account a certain locus, one can use A*, B*, ..., DPB1* as the allele name for both recipients and donors on this locus and the eplet mismatch computation will only take into account the loci filled in.

b. Not creating a file but generating a pandas.DataFrame

If one wants to generate a pandas.DataFrame directly, the output_path argument of simple_comparison can be set to None. The pandas.DataFrame will be returned by the function. Same goes for compute_epletic_load.

Exit codes:

- 55: an eplet did not match the regular expression '^\d+' (ABC, DR, DQ or DP) and it also did not match the regular
expression '^.[pqr]*(\d+)' (interlocus2) either.

Unit tests

Tested on Python 3.10.2 & Python 3.11.0.

platform win32 -- Python 3.10.2, pytest-7.2.0, pluggy-1.0.0
plugins: mypy-0.10.0
collected 19 items                                                                                                                                     

unit_tests_mypy.py ..                                                               [ 10%]
unit_tests_simple.py .                                                              [ 15%] 
pelc\__init__.py .                                                                  [ 21%] 
pelc\_unexpected_alleles.py .                                                       [ 26%] 
pelc\epitope_comparison.py .                                                        [ 31%] 
pelc\epitope_comparison_aux.py .                                                    [ 36%] 
pelc\output_type.py .                                                               [ 42%] 
tests\__init__.py .                                                                 [ 47%] 
tests\base_loading_for_tests.py .                                                   [ 52%] 
tests\test_epitope_mismatches.py ......                                             [ 78%]
tests\test_pelc.py ..                                                               [ 89%] 
tests\test_unexpected_alleles.py ..                                                 [100%]

platform win32 -- Python 3.11.0, pytest-7.2.0, pluggy-1.0.0
plugins: mypy-0.10.0
collected 19 items

unit_tests_mypy.py ..                                                               [ 10%]
unit_tests_simple.py .                                                              [ 15%]
pelc\__init__.py .                                                                  [ 21%]
pelc\_unexpected_alleles.py .                                                       [ 26%]
pelc\epitope_comparison.py .                                                        [ 31%]
pelc\epitope_comparison_aux.py .                                                    [ 36%]
pelc\output_type.py .                                                               [ 42%]
tests\__init__.py .                                                                 [ 47%]
tests\base_loading_for_tests.py .                                                   [ 52%]
tests\test_epitope_mismatches.py ......                                             [ 78%]
tests\test_pelc.py ..                                                               [ 89%]
tests\test_unexpected_alleles.py ..                                                 [100%]

About the source code

• Follows PEP8 Style Guidelines.
• All functions are unit-tested with pytest.
• All variables are correctly type-hinted, reviewed with static type checker mypy.
• All functions are documented with docstrings.

Useful links:

• Corresponding GitHub repository
• Corresponding PyPI page

Citation

If you use this software, please cite it as below.

• APA:

If you use this software, please cite it as below. Lhotte, R., Clichet, V., Usureau, C. & Taupin, J. (2022). 
Python Eplet Load Calculator (PELC) package (Version 0.4.1) [Computer software]. https://doi.org/10.5281/zenodo.7254809

• BibTeX:

@software{lhotte_romain_2022_7526198,
  author       = {Lhotte, Romain and
                  Clichet, Valentin and
                  Usureau, Cédric and
                  Taupin, Jean-Luc},
  title        = {Python Eplet Load Calculator},
  month        = oct,
  year         = 2022,
  publisher    = {Zenodo},
  version      = {0.4.1},
  doi          = {10.5281/zenodo.7526198},
}