Sign of Kukol Order Parameters

[TomPiggot]

While performing some analysis for the PS manuscript, I have noticed that the signs of the order parameters for the Kukol POPC results in the paper (Figure 2 which plots SCH and not -SCH; also referred to in Figure 4) are incorrect.

From my analysis of the simulation (https://zenodo.org/record/13393), I get:

beta 0.057708 -0.00331353
alpha 0.122723 0.136159
g3 -0.113621 -0.285188
g2 -0.140658
g1 0.237222 0.021846

These results now also make much more sense when comparing to the plotted results of related force fields (e.g. Berger, Poger, Chiu, etc.).

[markussmiettinen]

@mattijavanainen, could you confirm @TomPiggot's finding, as I think the original values were contributed by you?

http://nmrlipids.blogspot.com/2014/03/the-lipid-forcefield-comparison.html?showComment=1411160765055#c4481903312421738880

[mattijavanainen]

I calculated the values with something I wrote based on the template.c bundled with GROMACS 4.0. I also remember that my tool was giving out -S_CD so that for reporting S_CD, I reverted the signs when commenting them to the blog. Unfortunately, it seems that I forgot to do this reversal for the Kukol model (I trust Tom's results without further checking as they indeed match the values of other related models).

So should we replot Fig. 2 and submit a correction?

[markussmiettinen]

In addition to Fig. 2, also Fig. 4 would need checking. Kukol was at the bottom there.

[ohsOllila]

I think that it would be good to write a correction now. We could also include the issue related to the initial structures: http://nmrlipids.blogspot.com/2016/01/does-glycerol-backbone-structure-depend.html

Ideally, we could publish a updated version of the manuscript under the same reference, however, J. Phys. Chem. B is probably not advanced enough to do this. Easiest way would be probably to write a traditional correction with new figures and also update the figures in the version in the GitHub repository. I believe that the actual conclusions should not change, so this should not be very much work. However, I am currently very busy with other NMRlipids pojects and with few other things, so I am not sure when I would have time to this. If someone wants to start to work on this already now, it would be really good.

We also have to decide what is the policy for the authorship in the correction.