[gmx-users] Misunderstandings with g_rms

[Mark Abraham]

Jochen Hub wrote:
> DimitryASuplatov wrote:
>> Hello,
>>
>> I am sorry for this stupid question.
>>
>> Lets say I have an index file of all chain A atoms. I want to calculate
>> RMS of chain A _backbone_ only.
>>
>> I run g_rms. It asks two questions:
>> 1/ Select group for least squares fit
>> 2/ Select group for RMSD calculation
>>
>> To the best of my understanding, I should answer "BACKBONE" to the first
>> question and enter group number corresponding to chain A index file to
>> the second one.
> 
> In your case you need to make an index file with a group of the backbone
> of chain a, do something like
> 
> make_ndx -f x.gro
> chain a
> "ChA" & Backbone
> q
> 
> then run g_rms with -n
> 
> And give the ChainA-backbone group twice to g_rms.

It's not necessarily true that the group used for fitting should be the 
group for the RMSD calculation - that's why g_rms allows them to be 
different. People should consider their objective and choose 
appropriately in each case.

Mark