Best Regards
Elvis Martis
On Thu, 11 Jul 2019 at 10:35, Lod King <lodking407.gmail.com> wrote:
> Hi,
>
> I then tried the following based on the manual:
>
> > cluster C0 :1-42.CA clusters 3 epsilon 5.0 out test.dat summary
> avg.summary.dat
> >> I am sure what happened here but minpts 3 will lad to large number of
> clusters.
> and obtained a total cluster of 50. The first 10 are:
>
> #Cluster Frames Frac AvgDist Stdev Centroid AvgCDist
> 0 752 0.376 3.980 1.153 1255 10.563
> 1 341 0.171 2.364 0.643 1890 11.069
> 2 137 0.069 3.497 0.868 467 9.567
> 3 117 0.059 2.882 0.791 572 10.541
> 4 90 0.045 3.704 1.086 122 9.680
> 5 74 0.037 3.385 0.959 796 10.716
> 6 63 0.032 2.320 0.668 878 11.336
> 7 48 0.024 3.206 1.137 312 10.564
> 8 40 0.020 3.487 0.845 752 10.079
> 9 38 0.019 3.990 1.074 204 9.960
> 10 36 0.018 4.254 1.026 650 10.361
>
>
> I wondered,
>
> 1. how is 50 defined? By default?
> >> in DBSCAN we select the minpts and epsilon and the # of clusters are
> then decided by the algorithm.
> 2. why is cluster number started from 0?
> >> yes. the most populated cluster is 0 (ie. first)
> 3. Is there any way to plot them beside specifying "filename.agr"
> >> It depends on what you would lile to plot
> 4. Actually, how to read the data? what is centroid number defined here?
> > #Cluster Frames Frac AvgDist Stdev Centroid AvgCDist the
> manual has defined all this terms in good detail.
> On Wed, Jul 10, 2019 at 8:09 PM Elvis Martis <elvis_bcp.elvismartis.in>
> wrote:
>
> > I guess there must be some typo or an extra \, just recheck all commands.
> > Moreover, the AMBER Manual mentions a method on how to select an epsilon
> > and minpts, you might want to check that once.
> > On Thursday, July 11, 2019, Lod King <lodking407.gmail.com> wrote:
> >
> > > Hi Amber
> > >
> > > I saw a workshop tutorial online for Clustering and tried to follow the
> > > command :
> > >
> > > *>parm abc.prmtop*
> > >
> > > *>trajin abc.dcd*
> > > *>cluster c0 \ dbscan minpoints 25 epsilon 0.9 sievetoframe \ rms
> > :1-42.CA
> > > \ sieve 2000 random \ out cnumvtime.dat \ sil Sil \ summary
> summary.dat \
> > > info info.dat\ cpopvtime cpopvtime.agr normframe \ repout rep repfmt
> pdb
> > \
> > > singlerepout siglerrep.nc <http://siglerrep.nc> singlerepfmt netcdf \
> > > avgout Avg avgfmt restart*
> > >
> > > but got the following message:
> > >
> > > CLUSTER: Using coords dataset _DEFAULTCRD_, clustering using RMSD
> > (mask
> > > [:1-42.CA]) best-fit
> > > DBSCAN:
> > > Minimum pts to form cluster= 25
> > > Cluster distance criterion= 0.900
> > > Sieved frames will only be added back if they are within
> > > 0.900 of a frame in an existing cluster.
> > > (This option is more accurate and will identify sieved
> > > frames as noise but is slower.)
> > > Initial clustering will be randomly sieved (with value 2000).
> > > Only non-sieved frames will be used to calc within-cluster average.
> > > Cluster # vs time will be written to cnumvtime.dat
> > > Cluster pop vs time will be written to cpopvtime.agr (normalized by
> > frame)
> > > Pairwise distance data set is 'c0[PWD]'
> > > Cluster information will be written to info.dat\
> > > Summary of cluster results will be written to summary.dat
> > > Frame silhouettes will be written to Sil.frame.dat, cluster silhouettes
> > > will be written to Sil.cluster.dat
> > > Silhouette calculation will use non-sieved frames ONLY.
> > > Representative frames will be chosen by closest distance to cluster
> > > centroid.
> > > Cluster representatives will be written to 1 traj (siglerrep.nc),
> format
> > > Amber NetCDF
> > > Cluster representatives will be written to separate trajectories,
> > > prefix (rep), format PDB
> > > Average structures for clusters will be written to Avg, format Amber
> > > Restart
> > > Error: [cluster] Not all arguments handled: [ \ \ \ \ \ \ \ \ \ \ ]
> > >
> > > Did I miss anything?
> > > _______________________________________________
> > > AMBER mailing list
> > > AMBER.ambermd.org
> > > http://lists.ambermd.org/mailman/listinfo/amber
> > >
> >
> >
> > --
> > Best Regards
> > Elvis Martis
> > Mumbai.
> > _______________________________________________
> > 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
>
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Wed Jul 10 2019 - 22:30:02 PDT
