Re: [gmx-users] Re: g_covar

Hi Gideon, Maybe you want the atomic matrix produced by g_covar -xpma? Cheers, Tsjerk 2011/8/16 גדעון לפידות : > Thanks for your replies. > I would like to clarify regarding my first questionn. I don't want a g_dist > matrix. I would like to get a covariance matrix where the values correspond >

Re: [gmx-users] Re: g_covar

There is similar question in the mail list: http://www.mail-archive.com/gmx-users@gromacs.org/msg41737.html Regards, Yuri 2011/8/16 גדעון לפידות > Thanks for your replies. > I would like to clarify regarding my first questionn. I don't want a g_dist > matrix. I would like to get a covariance ma

[gmx-users] Re: g_covar

Thanks for your replies. I would like to clarify regarding my first questionn. I don't want a g_dist matrix. I would like to get a covariance matrix where the values correspond to absolute distance and not dived into different dimensions. for example say I have a protein with 100 aa and I run g_cov

Re: [gmx-users] Re: g_covar

Hi, The covariances are defined as the second central moment, i.e. the mean square displacement about the mean. Thus to make the PCA interpretable straightforwardly, you'll need to calculate the fluctuations with respect to the average structure. This stands apart from fitting. The fit is perform

[gmx-users] Re: g_covar

Dear Dagoberto, Choosing the right protein structure (and also which part of the protein to align) depends a lot of your system. I wouldn't use an average structure for a system that's changing a lot, because it may not be representative. It's hard to provide a more elaborate answer without many d