namd-l: Re: ABF simulation, free energy to remove a DNA strand

Hi Charlie,

Depending on the length of the strands, the distance coordinates that you are using could be insufficient to describe the process, especially if there is a lot of twisting involved as well.

A while ago there was one ABF study on a related question: http://pubs.acs.org/doi/abs/10.1021/jp0762323

That was a very local calculation, which is probably the reason why it went well.

In the end you might have to decompose the process into several steps or subprocesses described by different coordinates. They might happen sequentially or concurrently.

Best,

Jerome

On 5 August 2014 08:54, Charles Whidborne <charles.whidborne@student.adelaide.edu.au> wrote:

Hi,

I am running simulations on DNA in which i have equilibrated it over so many nanoseconds and now i want to measure the free energy of separating the strands using adaptive biasing force (ABF) in NAMD, but due to my inexperience (and probably lack of understanding) i am having trouble running the simulation successfully. I have done the ABF tutorials but they dont seem to help much when I try to relate it to my problem.

In my NAMD config file i have:
colvars                 on
colvarsConfig        Distance.in

In my Distance.in file i have set:
colvarsTrajFrequency      200
colvarsRestartFrequency 2000

colvar {
   name COMDistance

   width 0.1
   lowerboundary 1.0        -Not sure if these two boundaries are set appropriately.
   upperboundary 32.0

   distance {
      group1 {
         atomnumbers { #'s for back bone of one strand }
      }
      group2 {
         atomnumbers { #'s for backbone of second strand }
      }
   }
}

abf {
   colvars COMDistance
   fullSamples   500
   hideJacobian
}

I have had a couple simulations result in the partial separation of the strands, but as far as i can tell, due to the wrapping around each other the strand then gets 'stuck' trying to be pulled around the second strand... Any help or advice would be great on this.

Thanks,

Charlie