1GMX-CLUSTER(1) GROMACS GMX-CLUSTER(1)
2
6 gmx-cluster - Cluster structures
7
9 gmx cluster [-f [<.xtc/.trr/...>]] [-s [<.tpr/.gro/...>]] [-n [<.ndx>]]
10 [-dm [<.xpm>]] [-om [<.xpm>]] [-o [<.xpm>]] [-g [<.log>]]
11 [-dist [<.xvg>]] [-ev [<.xvg>]] [-conv [<.xvg>]]
12 [-sz [<.xvg>]] [-tr [<.xpm>]] [-ntr [<.xvg>]]
13 [-clid [<.xvg>]] [-cl [<.xtc/.trr/...>]]
14 [-clndx [<.ndx>]] [-b <time>] [-e <time>] [-dt <time>]
15 [-tu <enum>] [-[no]w] [-xvg <enum>] [-[no]dista]
16 [-nlevels <int>] [-cutoff <real>] [-[no]fit]
17 [-max <real>] [-skip <int>] [-[no]av] [-wcl <int>]
18 [-nst <int>] [-rmsmin <real>] [-method <enum>]
19 [-minstruct <int>] [-[no]binary] [-M <int>] [-P <int>]
20 [-seed <int>] [-niter <int>] [-nrandom <int>]
21 [-kT <real>] [-[no]pbc]
22
24 gmx cluster can cluster structures using several different methods.
25 Distances between structures can be determined from a trajectory or
26 read from an .xpm matrix file with the -dm option. RMS deviation after
27 fitting or RMS deviation of atom-pair distances can be used to define
28 the distance between structures.
29 single linkage: add a structure to a cluster when its distance to any
31 element of the cluster is less than cutoff.
32 Jarvis Patrick: add a structure to a cluster when this structure and a
34 structure in the cluster have each other as neighbors and they have a
35 least P neighbors in common. The neighbors of a structure are the M
36 closest structures or all structures within cutoff.
37 Monte Carlo: reorder the RMSD matrix using Monte Carlo such that the
39 order of the frames is using the smallest possible increments. With
40 this it is possible to make a smooth animation going from one structure
41 to another with the largest possible (e.g.) RMSD between them, however
42 the intermediate steps should be as small as possible. Applications
43 could be to visualize a potential of mean force ensemble of simulations
44 or a pulling simulation. Obviously the user has to prepare the trajec‐
45 tory well (e.g. by not superimposing frames). The final result can be
46 inspect visually by looking at the matrix .xpm file, which should vary
47 smoothly from bottom to top.
48 diagonalization: diagonalize the RMSD matrix.
50 gromos: use algorithm as described in Daura et al. (Angew. Chem. Int.
52 Ed. 1999, 38, pp 236-240). Count number of neighbors using cut-off,
53 take structure with largest number of neighbors with all its neighbors
54 as cluster and eliminate it from the pool of clusters. Repeat for re‐
55 maining structures in pool.
56 When the clustering algorithm assigns each structure to exactly one
58 cluster (single linkage, Jarvis Patrick and gromos) and a trajectory
59 file is supplied, the structure with the smallest average distance to
60 the others or the average structure or all structures for each cluster
61 will be written to a trajectory file. When writing all structures, sep‐
62 arate numbered files are made for each cluster.
63 Two output files are always written:
65 • -o writes the RMSD values in the upper left half of the matrix and
67 a graphical depiction of the clusters in the lower right half When
68 -minstruct = 1 the graphical depiction is black when two struc‐
69 tures are in the same cluster. When -minstruct > 1 different col‐
70 ors will be used for each cluster.
71 • -g writes information on the options used and a detailed list of
73 all clusters and their members.
74 Additionally, a number of optional output files can be written:
76 • -dist writes the RMSD distribution.
78 • -ev writes the eigenvectors of the RMSD matrix diagonalization.
80 • -sz writes the cluster sizes.
82 • -tr writes a matrix of the number transitions between cluster
84 pairs.
85 • -ntr writes the total number of transitions to or from each clus‐
87 ter.
88 • -clid writes the cluster number as a function of time.
90 • -clndx writes the frame numbers corresponding to the clusters to
92 the specified index file to be read into trjconv.
93 • -cl writes average (with option -av) or central structure of each
95 cluster or writes numbered files with cluster members for a se‐
96 lected set of clusters (with option -wcl, depends on -nst and
97 -rmsmin). The center of a cluster is the structure with the small‐
98 est average RMSD from all other structures of the cluster.
99
101 Options to specify input files:
102 -f [<.xtc/.trr/...>] (traj.xtc) (Optional)
104 Trajectory: xtc trr cpt gro g96 pdb tng
105 -s [<.tpr/.gro/...>] (topol.tpr)
107 Structure+mass(db): tpr gro g96 pdb brk ent
108 -n [<.ndx>] (index.ndx) (Optional)
110 Index file
111 -dm [<.xpm>] (rmsd.xpm) (Optional)
113 X PixMap compatible matrix file
114 Options to specify output files:
116 -om [<.xpm>] (rmsd-raw.xpm)
118 X PixMap compatible matrix file
119 -o [<.xpm>] (rmsd-clust.xpm)
121 X PixMap compatible matrix file
122 -g [<.log>] (cluster.log)
124 Log file
125 -dist [<.xvg>] (rmsd-dist.xvg) (Optional)
127 xvgr/xmgr file
128 -ev [<.xvg>] (rmsd-eig.xvg) (Optional)
130 xvgr/xmgr file
131 -conv [<.xvg>] (mc-conv.xvg) (Optional)
133 xvgr/xmgr file
134 -sz [<.xvg>] (clust-size.xvg) (Optional)
136 xvgr/xmgr file
137 -tr [<.xpm>] (clust-trans.xpm) (Optional)
139 X PixMap compatible matrix file
140 -ntr [<.xvg>] (clust-trans.xvg) (Optional)
142 xvgr/xmgr file
143 -clid [<.xvg>] (clust-id.xvg) (Optional)
145 xvgr/xmgr file
146 -cl [<.xtc/.trr/...>] (clusters.pdb) (Optional)
148 Trajectory: xtc trr cpt gro g96 pdb tng
149 -clndx [<.ndx>] (clusters.ndx) (Optional)
151 Index file
152 Other options:
154 -b <time> (0)
156 Time of first frame to read from trajectory (default unit ps)
157 -e <time> (0)
159 Time of last frame to read from trajectory (default unit ps)
160 -dt <time> (0)
162 Only use frame when t MOD dt = first time (default unit ps)
163 -tu <enum> (ps)
165 Unit for time values: fs, ps, ns, us, ms, s
166 -[no]w (no)
168 View output .xvg, .xpm, .eps and .pdb files
169 -xvg <enum> (xmgrace)
171 xvg plot formatting: xmgrace, xmgr, none
172 -[no]dista (no)
174 Use RMSD of distances instead of RMS deviation
175 -nlevels <int> (40)
177 Discretize RMSD matrix in this number of levels
178 -cutoff <real> (0.1)
180 RMSD cut-off (nm) for two structures to be neighbor
181 -[no]fit (yes)
183 Use least squares fitting before RMSD calculation
184 -max <real> (-1)
186 Maximum level in RMSD matrix
187 -skip <int> (1)
189 Only analyze every nr-th frame
190 -[no]av (no)
192 Write average instead of middle structure for each cluster
193 -wcl <int> (0)
195 Write the structures for this number of clusters to numbered
196 files
197 -nst <int> (1)
199 Only write all structures if more than this number of structures
200 per cluster
201 -rmsmin <real> (0)
203 minimum rms difference with rest of cluster for writing struc‐
204 tures
205 -method <enum> (linkage)
207 Method for cluster determination: linkage, jarvis-patrick,
208 monte-carlo, diagonalization, gromos
209 -minstruct <int> (1)
211 Minimum number of structures in cluster for coloring in the .xpm
212 file
213 -[no]binary (no)
215 Treat the RMSD matrix as consisting of 0 and 1, where the
216 cut-off is given by -cutoff
217 -M <int> (10)
219 Number of nearest neighbors considered for Jarvis-Patrick algo‐
220 rithm, 0 is use cutoff
221 -P <int> (3)
223 Number of identical nearest neighbors required to form a cluster
224 -seed <int> (0)
226 Random number seed for Monte Carlo clustering algorithm (0 means
227 generate)
228 -niter <int> (10000)
230 Number of iterations for MC
231 -nrandom <int> (0)
233 The first iterations for MC may be done complete random, to
234 shuffle the frames
235 -kT <real> (0.001)
237 Boltzmann weighting factor for Monte Carlo optimization (zero
238 turns off uphill steps)
239 -[no]pbc (yes)
241 PBC check
242
244 gmx(1)
245 More information about GROMACS is available at <‐
247 http://www.gromacs.org/>.
248
250 2022, GROMACS development team
2512022.2 Jun 16, 2022 GMX-CLUSTER(1)