Re: [AMBER] rms2d analysis

From: Daniel Roe <daniel.r.roe.gmail.com>
Date: Tue, 2 Apr 2019 15:43:54 -0400

These RMSD values are extremely small - all < 1 Ang. I don't think you
need cluster analysis here - everything is really close.

If you were to run it on a larger set you'd probably want to view the
output as a "map" - in that case it's better to write in gnuplot
format.

Depending on the type of clustering, I generally jump right to cluster
analysis on a subset of my trajectory, save the pairwise distance
matrix, and play around with the algorithm settings (e.g. number of
clusters/epsilon for hierarchical agglomerative, minpoints and epsilon
for dbscan, etc) until I get reasonable values for pseudo-F and DBI.
Then I scale things up.

Clustering is very much an art form. Be prepared for lots of trial and
error. I recommend checking out *the* classic MD clustering paper from
Shao, Cheatham et. al if you havent already:
https://pubs.acs.org/doi/abs/10.1021/ct700119m

-Dan

On Tue, Apr 2, 2019 at 2:38 PM Debarati DasGupta
<debarati_dasgupta.hotmail.com> wrote:
>
> Hi Users,
>
> I was reading through some mailing list archives whereby people have suggested using the "rms2d" tool of cpptraj to analyse the conformational space diversity before even going to clustering.
> What I could not understand from the output of the rms2d tool i show to interpret it and which rms value can be used as a threshold to identify structurally similar structures versus different conformations.
> I just ran it on 10 frames of my simulation and I get an output like this. I am using a GLUTAMATE "O" atom as it is in one of my main hotspot sites.
>
> #Frame 1 2 3 4 5 6 7 8 9 10
> 1 0.000 0.062 0.102 0.118 0.104 0.099 0.063 0.081 0.066 0.067
> 2 0.062 0.000 0.124 0.097 0.086 0.062 0.057 0.061 0.080 0.033
> 3 0.102 0.124 0.000 0.182 0.124 0.145 0.085 0.129 0.079 0.108
> 4 0.118 0.097 0.182 0.000 0.100 0.061 0.124 0.077 0.115 0.105
> 5 0.104 0.086 0.124 0.100 0.000 0.069 0.090 0.048 0.070 0.075
> 6 0.099 0.062 0.145 0.061 0.069 0.000 0.078 0.055 0.085 0.058
> 7 0.063 0.057 0.085 0.124 0.090 0.078 0.000 0.077 0.056 0.034
> 8 0.081 0.061 0.129 0.077 0.048 0.055 0.077 0.000 0.061 0.056
> 9 0.066 0.080 0.079 0.115 0.070 0.085 0.056 0.061 0.000 0.065
> 10 0.067 0.033 0.108 0.105 0.075 0.058 0.034 0.056 0.065 0.000
> ~
> ~
>
>
> Debarati
> _______________________________________________
> AMBER mailing list
> AMBER.ambermd.org
> http://lists.ambermd.org/mailman/listinfo/amber

_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Tue Apr 02 2019 - 13:00:02 PDT
Custom Search