[gmx-users] order parameters of partially unsaturated chains

[JAVIER CEREZO BASTIDA]

Hello Angel.

I found a detailed description of the order parameter conventions to  
describe molecular axis in a paper from Heller et al (JPC, 97,  
8343-8360), including the case of unsaturated carbons. It should be  
equivalent to the implementation in GROMAS.

In practice, I don't know the best way to do it, but your proposal  
sounds reasonable.

Javier

Ángel Piñeiro <angel.pineiro at usc.es> escribió:

> Dear all
> I wonder which is the best way to determine the deuterium order
> parameters of partially unsaturated chains like the oleoyl chain of
> POPC, more specifically I wonder if I should use the -unsat flag in
> g_order for this case. I checked several papers on unsaturated lipids,
> see for instance:
> 1.- http://persweb.wabash.edu/facstaff/FELLERS/publications/jpc_01.pdf
> 2.-
> http://www.ncbi.nlm.nih.gov/pmc/articles/PMC1225825/pdf/biophysj00084-0102.pdf
> 3.-
> http://www.sciencedirect.com/science?_ob=MImg&_imagekey=B94RW-4V8X9MG-P-1&_cdi=56421&_user=10&_pii=S0006349597782596&_coverDate=11%2F30%2F1997&_sk=%23TOC%2356421%231997%23999269994%23800076%23FLP%23display%23Volume_73,_Issue_5,_Pages_2249-2849_%28November_1997%29%23tagged%23Volume%23first%3D73%23Issue%23first%3D5%23date%23%28November_1997%29%23&view=c&_gw=y&wchp=dGLbVlz-zSkzV&md5=778ad27c9fc1e0b3874b178e36451cb4&ie=/sdarticle.pdf
> 4.- http://pubs.acs.org/doi/abs/10.1021/ja903529f
> 5.- http://www.ibpc.fr/UMR7099/pdf/Warschawski05-1.pdf
>
> but it seems that nobody mentions anything about a correction for the
> unsaturated carbons. Then I went to the code, I think this is the
> important part:
>
> if (bUnsat) {
>   /* Using convention for unsaturated carbons */
>   /* first get Sz, the vector from Cn to Cn+1 */
>   rvec_sub(x1[a[index[i+1]+j]], x1[a[index[i]+j]], dist);
>   length = norm(dist);
>   check_length(length, a[index[i]+j], a[index[i+1]+j]);
>   svmul(1/length, dist, Sz);-
>
>   /* this is actually the cosine of the angle between the double bond
>      and axis, because Sz is normalized and the two other components of
>      the axis on the bilayer are zero */
>   sdbangle += acos(Sz[axis]);
> } else {
>   /* get vector dist(Cn-1,Cn+1) for tail atoms */
>   rvec_sub(x1[a[index[i+1]+j]], x1[a[index[i-1]+j]], dist);
>   length = norm(dist);      /* determine distance between two atoms */
>   check_length(length, a[index[i-1]+j], a[index[i+1]+j]);
>
>   svmul(1/length, dist, Sz);
>   /* Sz is now the molecular axis Sz, normalized and all that */
> }
>
> As far as I see the difference is that by using the -unsat option, the
> atoms Cn, Cn+1 are used to measure theta instead of the atoms Cn-1, Cn
> +1. I think this could be reasonable when the whole chain is unsaturated
> but I am not sure that it is correct to use the -unsat option when we
> have two unsat carbons in the middle of a larger chain.... The
> calculation could also be done for the whole chain without the -unsat
> option, then repeat it for the unsaturated carbons plus their first
> neighbors with the -unsat option and then replace the results obtained
> from the second calculation for the central carbons in the curve
> obtained without the -unsat option... as previously suggested in this
> list (see
> http://lists.gromacs.org/pipermail/gmx-users/2010-February/049072.html)
> but I am not sure that is the best... I would appreciate any comment.
>
> Cheers,
>
> Ángel.
>