PyPermm

permeability modeling made easy

About

PyPermm is a Python-based reimplementation of the PerMM library written by Andrei Lomize and Alexey Kovalenko. The original code can be found here. This code has been relicensed with the permission of the original authors.

If you use pypermm, please cite the original PerMM publication.

Installation

Simply run uv pip install pypermm, pixi add --pypi pypermm, or pip install pypermm.

Usage

PyPermm primarily exposes a single function, run_permm, which takes a list of atomic symbols and a list of atomic coordinates (in Å).

run_permm returns a dictionary with a lot of useful computed properties, including:

• asatot, the total accessible surface area of the heavy atoms (in Å**2).
• E_bind, the minimum insertion energy (in kcal/mol), representing optimal binding to the membrane interface.
• logP_BLM, the predicted intrinsic permeability for black lipid membrances
• logP_plasma, the predicted intrinsic permeability for plasma membranes
• logP_BBB, the predicted intrinsic permeability for the blood–brain barrier
• logP_Caco2, the predicted intrinsic permeability for Caco-2 membranes
• logP_PAMPA, the predicted intrinsic permeability for PAMPA membranes
• z, an array of z-positions (in Å) through the membrane where energies were calculated
• energies, the insertion energy (in kcal/mol) at each z-position after optimizing orientation

Here's what running PyPermm looks like:

from pytest import approx
from pypermm.pypermm import run_permm

symbols = ["N", "C", "C", "C", ...]

xyz = [
    [5.04, 1.944, -8.324],
    [6.469, 2.092, -7.915],
    [7.431, 0.865, -8.072],
    [6.916, -0.391, -8.544],
    ...
]
       
result = run_permm(symbols, xyz)

assert result["asatot"] == approx(366.19, abs=0.01)
assert result["logP_BLM"] == approx(-6.85, abs=0.01)
assert result["logP_plasma"] == approx(-7.43, abs=0.01)
assert result["logP_BBB"] == approx(-5.31, abs=0.01)
assert result["logP_Caco2"] == approx(-5.23, abs=0.01)
assert result["logP_PAMPA"] == approx(-7.55, abs=0.01)
assert result["E_bind"] == approx(-0.297, abs=0.01)

Credits

This package was created with Cookiecutter and the jevandezande/uv-cookiecutter project template.