[gmx-users] create dihedrals' index giving the names of atoms from file

[Mark Abraham]

Claus Valka wrote:
> Dear Sir,
> 
> Thank you very much for your immediate responce.
> In order for my text file to succeed it should have
> the following format :
> 
> a C1 C2 C3 C4
> a C5 C6 C7 C7
> .
> .
> .
> q
> 
> The first line says that we want the atoms with names
> C1 C2 C3 C4 (depending on the names according to one's
> system).
> The last line is also very important because is says
> to the make_ndx command to save and quit (something
> also that I hadn't included in my file and as a result
> the command was running without exiting, yet with the
> 'a' in front I saw that it had created the indexes I
> wanted).
> 
> The output for the first three quadraplets is the
> following :
> 
> [ C1_C4_C7_C10 ]
>    1    4    7   10  456  459  462  465  911  914  917
>  920 1366 1369 1372
> 1375 1821 1824 1827 1830 2276 2279 2282 2285 2731 2734
> 2737 2740 3186 3189
> 3192 3195 3641 3644 3647 3650 4096 4099 4102 4105 4551
> 4554 4557 4560 5006
> 5009 5012 5015 5461 5464 5467 5470 5916 5919 5922 5925
> 6371 6374 6377 6380
> 6826 6829 6832 6835
> [ C4_C7_C10_C13 ]
>    4    7   10   13  459  462  465  468  914  917  920
>  923 1369 1372 1375
> 1378 1824 1827 1830 1833 2279 2282 2285 2288 2734 2737
> 2740 2743 3189 3192
> 3195 3198 3644 3647 3650 3653 4099 4102 4105 4108 4554
> 4557 4560 4563 5009
> 5012 5015 5018 5464 5467 5470 5473 5919 5922 5925 5928
> 6374 6377 6380 6383
> 6829 6832 6835 6838
> [ C7_C10_C13_C16 ]
>    7   10   13   16  462  465  468  471  917  920  923
>  926 1372 1375 1378
> 1381 1827 1830 1833 1836 2282 2285 2288 2291 2737 2740
> 2743 2746 3192 3195
> 3198 3201 3647 3650 3653 3656 4102 4105 4108 4111 4557
> 4560 4563 4566 5012
> 5015 5018 5021 5467 5470 5473 5476 5922 5925 5928 5931
> 6377 6380 6383 6386
> 6832 6835 6838 6841
> 
>>From the online gromacs manual the output I think
> should be :
> [ C1_C4_C7_C10 ]
>    1    4    7   10 
>    456  459  462  465
>    911  914  917  920
>    ...
>    ... and so on
> Is this really necessary?

Index files do not use whitespace to delimit subgroups of indices. 
You're getting 64 indices because you have 16 identical polymers and 
you're asking make_ndx for a set than spans across residues. I expect 
that in asking for "quadruplets" g_angle requires that they be in 
separate index groups, but I've never used g_angle. Try your above input 
file after adding commands to restrict the atom selection to only one of 
the polymers and then try g_angle again.

Mark

> I tried to calculate the dihedral distribution for the
> first quadraplet and it was calculated without any
> error.Can I choose all the quadraplets after the
> command :
> 
> g_angle -f traj.xtc -s topol.tpr -n index_created.ndx
> -b 0 -e (last frame) -type dihedral -od output.xvg
> 
> Thank you once again,
> Nikos