Re: [EXTERNAL] Re: constant pulling velocity in SMD

From: JC Gumbart (gumbart_at_ks.uiuc.edu)
Date: Fri Apr 27 2012 - 12:30:53 CDT

Thanks. It would have helped to also see your input PDB and SMD files (mineq-03.restart.pdb and smd01.ref) but I can already tell what's going on from this line in the log:

SMD 0 0.163 -0.393 18 0 0 -10491.3

At time zero, there should be no force on the ion. The fact that it's so large means the starting coordinate and the reference coordinate are in strong disagreement. Visually compare the two files to determine where you actually want to start the ion. You may need to restrain its position during equilibration.

On Apr 27, 2012, at 11:18 AM, Jiao, Dian NMN (-EXP) wrote:

> See attached conf and log files. The dcd is too big for attachment. I just ran another 5ps simulation with the SMDvel 0.001 which is supposed to be 100 times faster. However, the ion acts pretty much the same, speed wise. Travels about 20 A in 150 fs.
>
> From: JC Gumbart <gumbart_at_ks.uiuc.edu>
> Date: Thu, 26 Apr 2012 23:21:25 -0500
> To: Dian Jiao <dnjiao_at_sandia.gov>
> Cc: "namd-l_at_ks.uiuc.edu" <namd-l_at_ks.uiuc.edu>
> Subject: Re: [EXTERNAL] Re: namd-l: constant pulling velocity in SMD
>
> I can think of a number of possible reasons. Would it be possible for you to make your input and output files available on the web so I can take a look?
>
> On Apr 26, 2012, at 11:10 PM, Jiao, Dian NMN (-EXP) wrote:
>
>> This is exactly the tutorial I was following. The parameters were set the same with SMDvel 0.00001 and SMDk 5. I doublechecked, I used 0.00001 for the velocity which is 10A/ns. Even if it is 100 A/ns, isn't it odd for the ion to travel ~10 A in 50 fs (approximately 20 micro meter/ns)? I think the force constant I use here is pretty small and hence shouldn't have a dramatic affect on the speed of the ion either.
>>
>> Dian
>> From: JC Gumbart [gumbart_at_ks.uiuc.edu]
>> Sent: Thursday, April 26, 2012 5:15 PM
>> To: Jiao, Dian NMN (-EXP)
>> Cc: namd-l_at_ks.uiuc.edu
>> Subject: [EXTERNAL] Re: namd-l: constant pulling velocity in SMD
>>
>> The velocity is in A/timestep. Assuming you're using a 1 fs timestep, that means you're pulling at the rate of 100 A/ns, 10x what you expect.
>>
>> How well SMD controls the speed depends on the size of the force constant. See the tutorial Forcing Substrates Through Channels here: http://www.ks.uiuc.edu/Training/Tutorials/#freeenergymethods
>>
>>
>> On Apr 26, 2012, at 11:33 AM, Jiao, Dian NMN (-EXP) wrote:
>>
>>> Hi, I was trying to pull this metal ion through an ion channel along Z-axis in SMD simulation. SMDvel was set at 0.0001, which means it will travel around 10 A in 1ns. However, my ion travels a whole lot faster. It is first placed near the top of the channel (extracellular surface), it goes from 15.233 to 7.753 (z direction) in merely 50 fs. It doesn't slow down until it hits this spot where the binding between the ion and the lumen facing residues (5 Glu) is very strong. This position is fairly close to the bottom of the channel (intracellular surface). The ion seems to be stablized there and never gets past this position throughout 5ps simulation. Isn't SMDvel supposed to control the speed of the ion moving? Why is it still going so fast? What do I do to make it easy for the ion to escape the binding of that glutamic ring and keeps going down? Thanks in advance.
>>> Dian
>>
>>
>
> <smd01.log><smd01.conf>

This archive was generated by hypermail 2.1.6 : Tue Dec 31 2013 - 23:21:55 CST