mpharrigan

#!/bin/bash 

#SBATCH --partition=normal
#SBATCH --job-name=ip_engine
#SBATCH --output=ip_engine.out
#SBATCH --time=7-00:00:00
#SBATCH --time-min=02:00:00
#SBATCH --qos=normal
#SBATCH --nodes=1
#SBATCH --mem=64000

mpharrigan

set mol [mol new {traj_fn}  waitfor all]
mol addfile {top_fn} waitfor all

# Open data file
set sel [atomselect $mol all]
set nf [molinfo $mol get numframes]
set fp [open {dat_fn} r]
set line ""

# Each line of the data file corresponds to a frame

mpharrigan

alias ss='squeue --format="%.7i %9P %9j  %9u %.13L %.13l %.16S   %.2t %.5D %.5C  %15R " --sort="t,-S"'
alias ssr='ss --states=R'
alias ssp='ss --states=PD'

mpharrigan

ATOM  14902  C12 MY      1     -25.136  10.163   9.358  1.00  0.00      MEMB    
ATOM  14942  C13 MY      1     -24.372   9.630   8.213  1.00  0.00      MEMB    
ATOM  14945  C14 MY      1     -25.074   9.701   6.922  1.00  0.00      MEMB    
ATOM  14948  C15 MY      1     -25.460   8.518   6.468  1.00  0.00      MEMB    
ATOM  14951  C16 MY      1     -26.583   8.386   5.816  1.00  0.00      MEMB    
ATOM  14954  C17 MY      1     -27.221   6.099   4.694  1.00  0.00      MEMB    
ATOM  14957  C18 MY      1     -28.077   5.983   3.677  1.00  0.00      MEMB    
ATOM  14960  C19 MY      1     -28.644   4.770   3.859  1.00  0.00      MEMB    
ATOM  14963 C110 MY      1     -27.454   3.698   4.032  1.00  0.00      MEMB    
ATOM  14966 C111 MY      1     -26.798   3.485   2.647  1.00  0.00      MEMB    

mpharrigan

#!/bin/bash 
 
#SBATCH --partition=long # or short, defq
#SBATCH --job-name=whatever
#SBATCH --output=whatever.%j.out
#SBATCH --time=5-00:00:00
#SBATCH --nodes=1
#SBATCH --mail-type=ALL
#SBATCH --mail-user=email@email.com
#SBATCH --ntasks=12 # or however many

mpharrigan

from mdtraj.formats.dcd import DCDTrajectoryFile
from mdtraj.formats.prmtop import load_prmtop
km = KMedoids(...)

# "centers" is a pandas DataFrame with metadata about each trajectory
# from which we are pulling structures

def get_top(struc):
    return centers.query("structure == @struc")['top_fn'].iloc[0]

mpharrigan

from __future__ import print_function
import simtk.openmm as mm
import simtk.openmm.app as app

d = dict(base_dir="RUN2",
         state='state.xml')


with open("{base_dir}/system.xml".format(**d)) as f:
    system = mm.XmlSerializer.deserialize(f.read())

mpharrigan

# Basic usage / workflow
ds = dataset("**/*.dcd")

dihed = ds.derive('dihed', fmt='dir-npy')
dihed += DihedralFeaturizer().fit_transform(ds)

tica = dihed.derive('tica', fmt='hdf5')
tica += tICA().fit_transform(dihed)

clusters = tica.derive('clusters', fmt='hdf5')

mpharrigan

ds = dataset("**/*.dcd", top='struct.pdb')
diheds = ds.fit_transform_with(DihedralFeaturizer(), 'diheds')
ticas = diheds.fit_transform_with(tICA(), 'ticas')
kmeans = KMeans(n_clusters=8)
clusters = ticas.fit_transform_with(kmeans, 'clusters')
msm = MSM().fit(clusters)
dump(msm, 'msm.pkl')

mpharrigan

ffmpeg -i molecule.%5d.ppm -c:v libx264 -preset slow -crf 18 molecule.mkv

# -i       input filenames
# -c:v     codec (libx264 --> H.264)
# -preset  duh
# -crf     quality factor, lower = better, 18 = practically lossless
#          output filename

# For colors, make sure you're using yuv444 which should be the default in newer ffmpeg