[gmx-users] g_cluster (gromos method) creates clusters with members having RMSD greater than the cutoff to the cluster "middle"

[Christopher Neale]

Dear Users:

I have sorted out the issue and filed a redmine:

http://redmine.gromacs.org/issues/1688

________________________________________
From: gromacs.org_gmx-users-bounces at maillist.sys.kth.se <gromacs.org_gmx-users-bounces at maillist.sys.kth.se> on behalf of Christopher Neale <chris.neale at alum.utoronto.ca>
Sent: 12 February 2015 22:53
To: gromacs.org_gmx-users at maillist.sys.kth.se
Subject: [gmx-users] g_cluster (gromos method) creates clusters with members having RMSD greater than the cutoff to the cluster "middle"

Dear Users:

I have run g_cluster from gromacs 4.6.5 as follows:

g_cluster -s my.tpr -f tmp.xtc -method gromos -nofit -o rmsd-clust_nofit.xpm -g cluster_nofit.log -sz clust-size_nofit.xvg -cl clusters_nofit.pdb -n index.ndx -cutoff 0.275 -wcl 10 -cl finally.xtc

The top of cluster_nofit.log looks like this:
Using gromos method for clustering
Using RMSD cutoff 0.275 nm
The RMSD ranges from 0.0247008 to 5.03412 nm
Average RMSD is 0.45568
Number of structures for matrix 12391
Energy of the matrix is 4177.11 nm

Found 874 clusters

Writing middle structure for each cluster to finally.xtc
Writing all structures for the first 10 clusters with more than 1 structures to finally.xtc%03d.xtc

cl. | #st  rmsd | middle rmsd | cluster members
  1 | 4219  0.202 | 784700 .154 |  92900 103300 116100 125900 126500 133000 156600
...
    |           |             | 693900 694000 694200 694300 694600 694700 695700
...

So the frame at 694700 is in the first cluster.

However, when I use g_rms to look at the rmsd between frame at 784700 and 694700, I find that the RMSD is 0.29 nm.

I am confused because frame 784700 is listed as the "middle" (which I assume is the cluster centroid) so frame 694700 should not be included in this cluster.

When I put these two frames into a single .xtc with only 2 frames, g_cluster correctly puts them into two different clusters when I use a cutoff of 0.275 nm.

I understand that some of the members in a cluster can have a pairwise RMSD greater than the cutoff, but I thought that the "middle" of the cluster should have an RMSD less than the cutoff to all cluster members.

Can anybody help me figure out what I am missing?

Thank you,
Chris.

--
Gromacs Users mailing list

* Please search the archive at http://www.gromacs.org/Support/Mailing_Lists/GMX-Users_List before posting!

* Can't post? Read http://www.gromacs.org/Support/Mailing_Lists

* For (un)subscribe requests visit
https://maillist.sys.kth.se/mailman/listinfo/gromacs.org_gmx-users or send a mail to gmx-users-request at gromacs.org.