[gmx-users] g_cluster[output]

[Anton Feenstra]

Osmany Guirola Cruz wrote:
> Hi
>   I am doing cluster analysis of my simulation. I use gromos method, and 
> analyse the backbone, but i need that the output(clusters.pdb) was the entire 
> peptide. First time i did this
>   group for least squares fit       4(backbone)
>   group for output                      4(backbone)
> and i found 4 clusters.
> next time i did this 
>    group for least squares fit       4(backbone)
>   group for output                     1(protein) (this is really that i need)
> and i found 15 clusters 
> WHY? i just change "group for output"  i think that the result must be the 
> same(i don't change the trajectory files) 
> What can i do? 

It sure smells like a dirty bug to me, but I don't have the time to look
into it. There is an awful lot of index group and structures book-keeping
going on inside g_cluster... ;-)

Are you using the latest Gromacs version? If not, try the latest version.
If so, can you  reproduce it with (one of) the older ones?
Do you get the same error if you choose different groups, e.g. cluster on
Ca, and write backbone?

I think I remember doing something very similar with one of my peptide
simulations, so it must have worked. Alternatively, you could distill the
frame numbers from the g_cluster output (they are there somewhere), and
feed that to the 'frames.ndx' input to trjconv, that will produce an
output trajectory with only those frames with numbers from the index group.
It's probably a bit of puzzling and tinkering to get this working, and
you should be sure to carefully check if you get the right frames.

Finally, if you have more time on your hands than you can use otherwise,
you could dive into the source and try to find out what goes wrong...


-- 
Groetjes,

Anton
  _____________ _______________________________________________________
|             |                                                       |
|  _   _  ___,| K. Anton Feenstra                                     |
| / \ / \'| | | Dept. of Pharmacochem. - Vrije Universiteit Amsterdam |
|(   |   )| | | De Boelelaan 1083 - 1081 HV Amsterdam - Netherlands   |
| \_/ \_/ | | | Tel: +31 20 44 47608 - Fax: +31 20 44 47610           |
|             | Feenstra at chem.vu.nl - www.chem.vu.nl/~feenstra/       |
|             | "If You See Me Getting High, Knock Me Down"           |
|             | (Red Hot Chili Peppers)                               |
|_____________|_______________________________________________________|