GlyCONFORMER

GlyCONFORMERS is a Python package that assigns labels to glycan conformers, based on their torsion angle values, in order to differentiate various 3D-structures of one single glycan. It enables the automated assignment of the GlyCONFORMER string that was introduced in:

Grothaus et al. 2022, Exploration, Representation, and Rationalization of the Conformational Phase Space of N-Glycans, J. Chem. Inf. Model. 2022, 62, 20, 4992–5008 https://pubs.acs.org/doi/full/10.1021/acs.jcim.2c01049 for N-glycans.

Check the paper or tutorial for a detailed explanation of the GlyCONFORMER string generation. The workflow is not exclusivly designed for N-glycans, but can also be applyied to any other glycan type there is.

Installation

GlyCONFORMER is available via pip and docker:

pip

pip install GlyCONFORMER

Stable performance was only tested and verified up to python version 3.10. Pandas<=1.3.5 is required!

Recommendation - build via conda environment with following commands:

conda create -n name python=3.10 --no-default-packages
conda activate name
pip install GlyCONFORMER

Docker Image

This image includes:

• GlyCONFORMER pre-installed via pip install .
• Jupyter Notebook launched on container start
• Required dependencies like plumed, numpy, scikit-learn, etc.

docker pull grothaus/glyconformer:latest

docker run -p 8888:8888 grothaus/glyconformer:latest

Tutorial

The tutorial includes different N-glycan types and different complexity levels of how to obtain a GlyCONFORMER label string for custom glycan types and their recorded torsion angle values. The minimum example is given by the high-mannose type N-glycan M5, where all necessary information are read from the LIBRARY_GLYCANS folder by specifying the glycantype = "Man5":

M5 = Glyconformer(glycantype = "Man5")

Raw torsion angle values get converted to letters corresponding to their values and associate them to a certain minima of the free energy profile along that torsion angle.

The raw input data can be accessed via:

M5.colvar

and its translated variant via:

M5.binary

Additionally, the occurance of each conformer string is counted and outputted via:

M5.count

All this is performed in the background and final results can be outputted using various plotting functions, e.g.:

M5.distribution

Conformer labels are given on the x-axis and deviations from the most populated conformer indicated by explicit letters, where dots are used when no change in that torsion angle could be detected.

For more elaborate examples and a detailed explanation of how free energy profiles of each torsion angle are classified, check out the Tutorial_GlyCONFORMER.ipynb notebook.

Documentation

See documentation https://glyconformer.readthedocs.io/en/latest/index.html