[gmx-users] AFM/COM Pulling differences?

[chris.neale at utoronto.ca]

Disclaimer: I'm not an AFM experimentalist, this is based on what I have read.

I agree with Justin on the how, but have you thought carefully about  
the why? I doubt that you can complete simulations that pull as slowly  
as a real AFM machine: 0.5 um/s = moving 1 nm over 2 ms of simulation  
time. Even within the experimental technique, you will get different  
profiles with different pulling rates. I have seen the output from  
such simulations run by others and one gets many different results  
that are difficult to interpret. Note also that experimental AFM  
results are averaged over hundreds of repetitions in the work that I  
have read.

In any event, check your convergence by running many times from  
different random seeds and other startup conditions.

Just saying,
Chris.

-- original message --

C Johnson wrote:

[Hide Quoted Text]
/ Maybe I don't fully understand your objective, but if you're just  
interested in
/>/ secondary structure stability as a function of distance between  
two species, a />/ slow steered MD run could accomplish that; you  
don't need umbrella sampling.  If
/>/ you're trying to get free energies associated with different  
configurations,
/>/ however, then getting the PMF is a suitable method.
/
/ -Justin
/ I'd like to be able to have a single peptide with the N-terminus  
anchored and pull the C-terminus with constant
velocity so that a force vs displacement plot could be generated.
You can do that with any of the pulling methods except  
"constant_force," which,
per its name, is not a constant velocity pulling method :)  The  
outputs you will
get are force vs. time and displacement vs. time, but you can  
post-process that
to get the plot you want.

-Justin