I have aggregated the results from particles of varying permeability
(discussed here )
in a separate Google Sheet
so that the results can be quickly plotted and included in the working MSP.

➡️ This dataset shall be included here
in this repository and
I need to create a nicer plot for the MSP.

Use the simulated data set to train some (decision tree? ) ML model to classify 3 permeability classes , which shall eventually turn out to be equally classified by bilayer phases (judging by current results).

I have found that the noisy Friction profile
arises also from the fact that the umbrella potentials
that restrain the configuration in a too narrow region.
As a result, the particle decorrelates too quickly to
tell on the autocorrelation and calculate the friction well (likely explenation).

Using an updated setting revealed that the particles really exhibit a small peak in friction profile
at the place where they change from solvated to dry.
This is nicely in line with the work of Siewert and Berendsen,
who have found the same feature at that place too.

➡️ I have to update the friction profiles to use simulations that show their features.

Also, I should try using some filtering to reduce the noise in the profiles.
(e.g. from my previous work: from scipy.signal import butter, filtfilt)

currently the friction is in ps/nm^2, but would be much nicer if it were in ns/nm^2 (the results will not be in order of "thousands", but "ones")

The setting of properties in the main class
is a bit clumsy and can be improved in the following sense:

    @property
    def ensemble(self):
        r"""EnsembleData: Information on the sampled ensemble

        Returns
        -------
        ensemble : EnsembleData
        """
        return self.__ensemble

    @ensemble.setter
    def ensemble(self, ensemble):
        if not isinstance(ensemble, EnsembleData):
            raise TypeError('No known conversion from ' + str(type(ensemble)) +
                            'to EnsembleData')
        self.__ensemble = ensemble

from https://gitlab.com/gromacs/gromacs/-/blob/master/tests/physicalvalidation/physical_validation/data/simulation_data.py

this will come to SI.