calculate MSD and Lipid diffusion

[contact.life...@gmail.com]

Hi 

After trying hard to make a small script for calculating lateral lipid diffusion coefficient and MSD and searched different research artilces to get anything to start with but as I am new to this Python and MDAnalysis I was not been able to make a script . But I came with the following idea to calculate MSD 

import numpy as np

    r = np.sqrt(xdata**2 + ydata**2)

    diff = np.diff(r) #this calculates r(t + dt) - r(t)

    diff_sq = diff**2

    MSD = np.mean(diff_sq)

How to implement it in MD analysis to calculate MSD and diffusion coefficient through this. 

I tried to use leaflet finder to solve the problem but the following program is giving me an error "TypeError: 'generator' object has no attribute '__getitem__'

"

import sys

import MDAnalysis

import numpy as np

import MDAnalysis.coordinates

from MDAnalysis.analysis.leaflet import LeafletFinder

u= MDAnalysis.Universe("/path to psf.psf","/path to.dcd") 

headgroup_atoms= u.selectAtoms("name P*")

coord = headgroup_atoms.coordinates()

#print headgroup_atoms

#print coord

from MDAnalysis.analysis.distances import distance_array

adj = (distance_array(coord,coord))

import networkx as NX

leaflets=NX.connected_components(NX.Graph(adj))

A_lipids = headgroup_atoms[leaflets[0]].residues

B_lipids = headgroup_atoms[leaflets[1]].residues

if anyone can help me , I will be really grateful to you .... 

[Michael Lerner]

I don’t have any tips on the MDAnalysis side of things, but I did recently write some Python code to calculate MSD. I did a decent job speeding things up with numba (see the blog post: http://www.mglerner.com/blog/?p=52). However, in the end, you’ll be happier using an FFT-based algorithm, such as FCA (see the link in the comments).

Cheers,

-Michael

(Likely sent from my phone; please excuse the brevity.)

—

Michael G. Lerner

mglerner.com

--
You received this message because you are subscribed to the Google Groups "MDnalysis discussion" group.
To unsubscribe from this group and stop receiving emails from it, send an email to mdnalysis-discus...@googlegroups.com.
To post to this group, send email to mdnalysis-...@googlegroups.com.
Visit this group at http://groups.google.com/group/mdnalysis-discussion.
For more options, visit https://groups.google.com/d/optout.

[Tyler Reddy]

I have some unit-tested MDAnalysis / Python-based code for calculating lateral diffusion in simulations (repo: https://github.com/tylerjereddy/diffusion_analysis_MD_simulations docs: http://diffusion-analysis-md-simulations.readthedocs.org/en/latest/diffusion_analysis.html).

I should probably upload a sample IPython notebook of the calculation workflow, but the documentation is relatively thorough. Have to be a bit careful to watch the units and the way you select particles, etc.

I've used it for at least one publication now and a few others in the group as using it now as well.

Cheers,

Tyler

--

[Oliver Beckstein]

On 13 Jul, 2015, at 07:23, contact.life...@gmail.com wrote:

1) How can I get xmin, xmax, y min and Y max values and delta value for my trajectory

If your trajectory contains unitcell information, look at Timestep.dimensions:

import numpy as np

maxL = np.zeros(3)

for ts in u.trajectory:

L = ts.dimensions[:3]      # Lx, Ly, Lz

angles = ts.dimensions[3:]    # alpha, beta, gamma -- not needed

replace = (maxL < L)

maxL[replace] = L[replace]

print("L: {0}       maxL: {1}".format(L, maxL))

print("maximum dimensions: {}".format(maxL))

If you really need coordinates, use the same idea as above but instead of dimensions look at AtomGroup.bbox() ("bounding box" ). From the help:

Signature: u.atoms.bbox(**kwargs)
Docstring:
Return the bounding box of the selection.

The lengths A,B,C of the orthorhombic enclosing box are::

  L = AtomGroup.bbox()
  A,B,C = L[1] - L[0]

:Keywords:
  *pbc*
    ``True``: Move all atoms within the primary unit cell before calculation [``False``]

In code:

bb = u.atoms.bbox()

L = bb[1] - bb[0]

and then as above.

Hope that helps,

Oliver

--
Oliver Beckstein * orbe...@gmx.net
skype: orbeckst  * orbe...@gmail.com