Command-line analysis routines of Molecular Dynamics data with aggregating molecules.
Here are the steps to install cluster_tools in ~/programs, but you can of course change the installation directory.
-
Clone this repository on your computer.
mkdir -p ~/programs cd ~/programs git clone https://github.com/ahardiag/cluster_tools_md
-
Use a python environment to install the package on it, e.g. using
conda:conda create -n cluster_tools_md python=3.10
-
Build the package using
pip:cd ~/programs pip install ./cluster_tools_md
clustsize -h
eccentricity -h
multirdf -h
radialdens -hTo retrieve the same conda environment used during the development, use :
conda env create --file "~/programs/cluster_tools/cluster_tools_md.yml" -n cluster_tools_md
conda activate cluster_tools_mdComputes the size distributions of atomic clusters in gas phase for a set of MD trajectories.
This is equivalent to gmx clustsize in GROMACS, but provides more options and editable outputs.
usage: clustsize [-h] -i INPUTDIR -m RESNAME [-o OUTDIRNAME]
[-r REFFILE] [-f TRJFILE] [-b BEGIN] -e END [-v]
--mins MINS --eps EPS [--with-distance-matrix]
[--with-double-labelling] [--ignore-warnings]
[--pattern-dirs PATTERN_DIRS]
Find all aggregate clusters for a set of MD trajectories.
For selection commands, see MDAnalysis doc :
https://docs.mdanalysis.org/stable/documentation_pagses/selections.html
optional arguments:
-h, --help show this help message and exit
-i INPUTDIR, --inputdir INPUTDIR
directory with all trajectories to analyse
-m RESNAME, --resname RESNAME
Resname or molecule name. e.g: HC8
-o OUTDIRNAME, --outdirname OUTDIRNAME
Output Sub-directory name.
-r REFFILE, --reffile REFFILE
Commmon name of reference file in all simulation directories.
-f TRJFILE, --trjfile TRJFILE
Commmon name of trajectory file in all simulation directories.
-b BEGIN, --begin BEGIN
Start frame of analysis.
-e END, --end END Stop frame of analysis.
-v, --verbose Turn on for more details during the calculation.
--mins MINS Minimum number of points (atoms) to define a cluster.
--eps EPS Maximum distance between two points (atoms) in the same cluster.
--with-distance-matrix
Use total matrix distance to treat periodic boundary conditions. Slow for large systems!
--with-double-labelling
Use double labelling method to perform clustering analysis on two shifted set of positions. Allow to search for clusters close to the box boundaries.
--ignore-warnings Avoid stdout all warnings from using clustering on sparse matrix.
--pattern-dirs PATTERN_DIRS
String pattern to select a serie of data directories with similar pattern in the input directory.The script generates three CSV files located in OUTDIRNAME directory (by default : output) and these data are easily editables using, for instance, the DataFrame format from the pandas library.
-
cluster_histo.csv : size indices in columns and
$simulation$ key, aggregation number$n$ and size distribution$P(n)$ in rows. -
cluster_properties : time (in ns) in columns and
$simulation$ key, number of clusters$nclust$ and number of molecules in the biggest cluster$maxclust$ in rows. -
cluster_resids.csv : time (in ns) in columns and
$simulation$ key and$resid$ indices in rows. For a given simulation, time and resid index one have stored the label (positive or ) of the cluster where is found the molecule.
More insights of the output data is accesible in tests directory.
Run sequentially Radial Distribution Function analysis for a set of MD trajectories.
First you need to create a config file with some default parameters, in a file on the root location : ~/.runRDFrc:
[DefaultParameters]
BEGIN= 100 # Start time in ns for the analysis
END= 800
MAX= 20 # Maximum radial distance (in
# Angstrom)
SUB= 2 # Number of time intervals to
# divide the
# Trajectory total time
NBINS= 200 # Number of points for one RDF
# trace
PATH= ../Data_traj/ # Relative path to data from the
# current directory
EXCLUSION_BLOCK= (1,1) Then you need to specify the parameters you want to change for each analysis in a input file, e.g. parameters.in:
TASK SIM OUTDIRNAME FREQ SEL1 EXCLUSION_BLOCK
taks1 U01_3ch run0 1000 "resname U01 and name N12" (1,1)
task2 U02_3ch run0 1000 "resname U02 and name N4" (1,1)
Then you just have to run the main script in a directory where you want to store results:
multirdfCompute the eccentricity of each cluster found after using clustsize command.
usage: eccentricity [-h] -s SIM -i INPUTDIR -c CSVDIR [--nbins NBINS] [-m RESNAME] [-o OUTDIRNAME]
[--outfigdirname OUTFIGDIRNAME] [-r FILEREF] [-f FILETRJ]
[--method {double_loop,best_centering}] [--compound {residues,atoms}]
[--rename RENAME] [--in_memory] [-z SIZE [SIZE ...]] [--range] [-v]
Compute Spherical Radial Density around cluster COM.
options:
-h, --help show this help message and exit
-s SIM, --sim SIM Name of the Data trajectory.
-i INPUTDIR, --inputdir INPUTDIR
directory with all trajectories to analyse.
-c CSVDIR, --csvdir CSVDIR
directory with the dataframe `cluster_resids.csv`.
--nbins NBINS Number of bins of the density distribution.
-m RESNAME, --resname RESNAME
Resname or molecule name. e.g: HC8.
-o OUTDIRNAME, --outdirname OUTDIRNAME
Output Sub-directory name.
--outfigdirname OUTFIGDIRNAME
Output Sub-directory name for figures.
-r FILEREF, --fileref FILEREF
Commmon name of reference file in all
simulation directories.
-f FILETRJ, --filetrj FILETRJ
Commmon name of trajectory file in all
simulation directories.
--method {double_loop,best_centering}
Choose the method for computing the inertia tensor.
--compound {residues,atoms}
Choose the level at which is computed the inertia tensor.
`residues` is faster as it only consider the center of mass of the molecules.
--rename RENAME Set to True to consider sim name with no suffix `Data_`.
--in_memory Charge the whole trajectory in RAM memory.
Faster than reading from file.
-z SIZE [SIZE ...], --size SIZE [SIZE ...]
Choose specific size(s) of clusters.
--range Range of values provided by three values `start end step`by -z/--size option.
-v, --verbose Turn on for more details during the
calculation.Compute the radial density of an atom group around each cluster, depending on its size.
usage: radialdens [-h] -s SIM -z SIZE [SIZE ...] [--range] -i INPUTDIR -c CSVDIR [--nbins NBINS]
[-m RESNAME] [-o OUTDIRNAME] [--outfigdirname OUTFIGDIRNAME] [-r FILEREF]
[-f FILETRJ] [--rename RENAME] [--in_memory] [-v]
Compute Spherical Radial Density around cluster COM.
options:
-h, --help show this help message and exit
-s SIM, --sim SIM Name of the Data trajectory.
-z SIZE [SIZE ...], --size SIZE [SIZE ...]
Size of the cluster.
--range Range of values provided by three values `start end step`by -z/--size option.
-i INPUTDIR, --inputdir INPUTDIR
directory with all trajectories to analyse.
-c CSVDIR, --csvdir CSVDIR
directory with the dataframe `cluster_resids.csv`.
--nbins NBINS Number of bins of the density distribution.
-m RESNAME, --resname RESNAME
Resname or molecule name. e.g: HC8.
-o OUTDIRNAME, --outdirname OUTDIRNAME
Output Sub-directory name.
--outfigdirname OUTFIGDIRNAME
Output Sub-directory name for figures.
-r FILEREF, --fileref FILEREF
Commmon name of reference file in all
simulation directories.
-f FILETRJ, --filetrj FILETRJ
Commmon name of trajectory file in all
simulation directories.
--rename RENAME Set to True to consider sim name with no suffix `Data_`.
--in_memory Charge the whole trajectory in RAM memory.
Faster than reading from file.
-v, --verbose Turn on for more details during the
calculation.A set of scripts (test1.sh,test2.sh,...) are given as examples and control tests, it processes the data from the article cited in section Citation.
Run the two first tests
cd ~/programs/cluster_tools_md/tests/
./run_tests.sh test1 test2If you use cluster_tools in your research, please cite the following article:
[coming soon]