With that many frames and that many atoms in the mask you might be having
memory issues. You could try using a larger sieve, or maybe better to
reduce your mask to only heavy atoms, or only backbone.
On Sat, May 22, 2021, 9:43 AM Sruthi Sudhakar <sruthisudhakarraji.gmail.com>
wrote:
> Dear all,
>
>
> I have been trying to cluster my Gaussian Accelerated MD trajectory of a
> protein-nucleic acid system using the following script.
> parm noions.prmtop
>
> trajin gamd2.nc
> loadcrd gamd2.nc 1 last 5 clus_crd
>
> cluster hieragglo epsilon 3 clusters 10 averagelinkage rms :1-1362 mass
> sieve 10 out clus_out.dat summary clus_summary.dat info clus_info3.dat
> repout rep3 repfmt pdb
>
> The trajectory has 250000 frames (500ns dt =0.002 with frames saved for
> every 1000 steps). The clustering proceeds for 10-20% and gets killed.
> Could someone suggest any way out of this?
> Thanks in advance.
>
> Regards,
> Sruthi
>
>
> INPUT: Reading input from 'cluster.in'
> [parm noions.prmtop]
> Reading 'noions.prmtop' as Amber Topology
> Radius Set: modified Bondi radii (mbondi)
> [trajin gamd2.nc]
> Reading 'gamd2.nc' as Amber NetCDF
> [loadcrd gamd2.nc 1 last 5 clus_crd]
> Reading 'gamd2.nc' as Amber NetCDF
> Warning: COORDS data sets do not store times.
> Loading trajectory 'gamd2.nc' as 'clus_crd'
> ----- gamd2.nc (1-250000, 5) -----
> [cluster hieragglo epsilon 3 clusters 10 averagelinkage rms :1-1362 mass
> sieve 10 out clus_out.dat summary clus_summary.dat info clus_info3.dat
> repout rep3 repfmt pdb]
> CLUSTER: Using coords dataset _DEFAULTCRD_, clustering using RMSD (mask
> [:1-1362]), mass-weighted best-fit
> Hierarchical Agglomerative: 10 clusters, epsilon 3.000, average-linkage.
> Sieved frames will not be included in final cluster distance calculation.
> Initial clustering sieve value is 10 frames.
> Only non-sieved frames will be used to calc within-cluster average.
> Cluster # vs time will be written to clus_out.dat
> Pairwise distance data set is 'Cnum_00003[PWD]'
> Cluster information will be written to clus_info3.dat
> Summary of cluster results will be written to clus_summary.dat
> Representative frames will be chosen by closest distance to cluster
> centroid.
> Cluster representatives will be written to separate trajectories,
> prefix (rep3), format PDB
> Warning: One or more analyses requested creation of default COORDS DataSet.
> CREATECRD: Saving coordinates from Top noions.prmtop to "_DEFAULTCRD_"
> ---------- RUN BEGIN -------------------------------------------------
>
> PARAMETER FILES (1 total):
> 0: noions.prmtop, 27735 atoms, 1530 res, box: Orthogonal, 7 mol
>
> INPUT TRAJECTORIES (1 total):
> 0: 'gamd2.nc' is a NetCDF AMBER trajectory with coordinates, time, box,
> Parm noions.prmtop (Orthogonal box) (reading 250000 of 250000)
> Coordinate processing will occur on 250000 frames.
>
> BEGIN TRAJECTORY PROCESSING:
> .....................................................
> ACTION SETUP FOR PARM 'noions.prmtop' (1 actions):
> 0: [createcrd _DEFAULTCRD_]
> Warning: COORDS data sets do not store times.
> Estimated memory usage (250000 frames): 83.211 GB
> ----- gamd2.nc (1-250000, 1) -----
> 0% 10% Killed
> _______________________________________________
> 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 Sat May 22 2021 - 08:30:03 PDT
