Re: ion escapes the channel during US simulation

From: Vermaas, Joshua (Joshua.Vermaas_at_nrel.gov)
Date: Thu Sep 01 2016 - 10:27:16 CDT

You are exactly correct in your interpretation of the colvar traj file,
but look where the colvar is at t=0. 9.74 *10^0 = 9.74. This is
naturally nowhere near the center you specified (132ish if I remember
correctly), so the colvar does what you tell it to do, and will push the
ion hard out of the channel to try and get the relative distance between
the ion and the channel to be 132 Angstroms in Z. This is why I
recommended you check the value of the colvar in VMD, which would have
told you that your centers are way off. The range of the collective
variable you should be exploring is probably something like 0-20 or
something like that, not stuff with these huge numbers.

-Josh

On 09/01/2016 09:17 AM, Olya Kravchenko wrote:
> Hi Josh,
>
> here is part of the .colvars.traj :
>
> 0 9.74902505517626e+00 0.00000000000000e+00 2.36501949889647e+01
> 100 1.05933088334863e+01 -3.29257204531281e+01 2.34813382333027e+01
> 200 1.09898368810024e+01 2.48431274146785e+02 2.34020326237995e+01
> 300 1.15130838051553e+01 -6.27930827442560e+01 2.32973832389689e+01
> 400 1.21147065307773e+01 -3.18638108785592e+01 2.31770586938445e+01
> 500 1.24867184768613e+01 1.58916284983259e+01 2.31026563046277e+01
> 600 1.31679543764748e+01 -1.20348764469353e+02 2.29664091247050e+01
> 700 1.39480460101585e+01 -4.19770289782558e+02 2.28103907979683e+01
> 800 1.47348780785561e+01 2.86632022042336e+02 2.26530243842888e+01
> 900 1.57315242933076e+01 5.99118384895983e+00 2.24536951413385e+01
> 1000 1.72287470089526e+01 4.58350573804379e+01 2.21542505982095e+01
> 1100 1.85251491118100e+01 -8.84642207258101e+01 2.18949701776380e+01
> 1200 1.94197770183663e+01 -6.23359178803364e+02 2.17160445963267e+01
> 1300 2.04581750243364e+01 3.34780129887652e+02 2.15083649951327e+01
> 1400 2.12922216706170e+01 -6.51671251034084e+01 2.13415556658766e+01
> 1500 2.27235764695745e+01 -1.58549929307424e+02 2.10552847060851e+01
> 1600 2.39084699531492e+01 3.15359678867356e+02 2.08183060093702e+01
> 1700 2.51314794853496e+01 1.67834568689645e+02 2.05737041029301e+01
> 1800 2.65028866070947e+01 8.49034008381802e+01 2.02994226785811e+01
>
>
> etc., in the end it looks like this:
>
> 509500 4.80237392589402e+01 -2.05971807923954e+02 1.59952521482120e+01
> 509600 4.79936675585035e+01 2.14449676372837e+01 1.60012664882993e+01
> 509700 4.79849852047113e+01 4.37927743569792e+02 1.60030029590577e+01
> 509800 4.79780077449055e+01 -2.41805488069026e+02 1.60043984510189e+01
> 509900 4.78868202340681e+01 5.53295692378051e+02 1.60226359531864e+01
> 510000 4.78683479038526e+01 -1.83869110846615e+02 1.60263304192295e+01
> 510100 4.79570223172412e+01 -3.70804106749638e+02 1.60085955365518e+01
>
> If I understand it correctly, the second column gives the value of
> the defined colvar (which in my case is distanceZ), so I would expect
> it to be somewhere near the "centers", which in this case is at 132A.
> Does it sound reasonable?
>
> Another thing that happens during these simulation (for each window)
> is that the residues closest to the entrance of the channel sort of
> follow the particle as it leave the cell and elongate the chain they
> belong to.
>
> Olga
>
> On Thu, Sep 1, 2016 at 10:44 AM, Vermaas, Joshua
> <Joshua.Vermaas_at_nrel.gov> wrote:
>> What are the values in the colvar file? Steadily increasing is fine if
>> they are going towards the target.
>> -Josh
>>
>> On 09/01/2016 08:06 AM, Olya Kravchenko wrote:
>>> Hi Amy,
>>>
>>> thank you very much for your comments. I believe that atom numbers are
>>> correct. The values in the colvar file indeed look strange, they are
>>> steadily increasing (I'd expect them to oscillate around certain
>>> value?). I will try to implement forceNoPBC.
>>>
>>> Olga
>>>
>>>
>>> On Wed, Aug 31, 2016 at 11:40 AM, Amy Rice <arice3_at_hawk.iit.edu> wrote:
>>>> Hi Olga,
>>>> A few things I would look at:
>>>> 1) This may be obvious, but have you verified that the atom numbers
>>>> (especially for the ion) are correct? I had an issue in the past where the
>>>> atom numbers of the original pdb I used to set up the system differed from
>>>> the numbering in the NAMD files which caused me to pull on the wrong atom.
>>>> Another possible point of confusion is that VMD uses 0 indexing for atoms
>>>> whereas NAMD starts the count at 1. Just a few simple things to check, as
>>>> they have caused me some confusion in the past with umbrella sampling!
>>>> 2) The next thing I would check is that the distanceZ colvar values match up
>>>> with what you think they should be. Take a few steps in the trajectory and
>>>> verify that the colvar value printed in the colvar.traj file seem reasonable
>>>> based on visualizing the trajectory. This could hint if you perhaps have the
>>>> center or groups specified incorrectly.
>>>> 3) Finally, you can try turning "forceNoPBC" on. I believe if you don't
>>>> specify this, distances are calculated according to the minimum image
>>>> distance, rather than the proper distance in the unit cell. I believe this
>>>> is probably the issue you are having, since you note that the ion is being
>>>> pulled across the periodic boundary, and faster when the force constant is
>>>> higher. Again, see point 2 because if this is the case, the values in the
>>>> colvar.traj file will probably not match with your expectation based on
>>>> visualizing the trajectory.
>>>>
>>>> Hope this helps,
>>>> - Amy
>>>>
>>>> On Tue, Aug 30, 2016 at 4:17 PM, Olya Kravchenko <ovkrav_at_gmail.com> wrote:
>>>>> Hi all,
>>>>>
>>>>> I try to run Umbrella Sampling with harmonic restraints on a ion
>>>>> permeation through the pore in a spherical protein. What I see is that
>>>>> ion is escaping through the entrance (while it supposed to either
>>>>> oscillate around the same spot or proceed towards the next binding
>>>>> site; this happened for the first three windows).
>>>>>
>>>>> I assumed that my force constant is not high enough to hold the ion
>>>>> near the desired position, so I experimented with the force constant
>>>>> and increased it up to 10 times, but what I see is an even faster
>>>>> escape. The ion moves so fast that it leaves the simulation cell and
>>>>> enters from the other side of the simulation cell and tries to
>>>>> approach the channel at the exit. The higher the force constant the
>>>>> faster it moves.
>>>>>
>>>>> I've added upperboundary and lowerboundary to no avail.
>>>>>
>>>>> Here is my current colvar input file. Am I making a mistake somewhere?
>>>>> Is there a way to keep the ion within desired range?
>>>>>
>>>>>
>>>>> colvarsTrajFrequency 100
>>>>> colvarsRestartFrequency 100
>>>>>
>>>>>
>>>>> colvar {
>>>>>
>>>>> width 1.0
>>>>> lowerboundary 108.0
>>>>> upperboundary 136.0
>>>>> outputSystemForce yes
>>>>> outputAppliedForce yes
>>>>> name ferritin
>>>>> distanceZ {
>>>>> main {
>>>>> atomnumbers { 291491 }
>>>>> }
>>>>> ref {
>>>>> atomnumbers {
>>>>> 5 19 30 47 63 87 104 118 139 156 173 185 196 211 221 231 250 264 288
>>>>> 305 324 338 357 372 391 412 422 433 454 470 491 510 521 538 549 570
>>>>> 591 611 623 647 659 671 687 697 716 738 752 772 782 804 825 845 864
>>>>> 881 898 909 926 941 956 980 995 1012 1022 1037 1059 1078 1095 1117
>>>>> 1136 1153 1167 1184 1208 1215 1222 1246 1265 1285 1304 1321 1333 1352
>>>>> 1369 1393 1405 1417 1428 1440 1452 1476 1491 1502 1509 1528 1542 1552
>>>>> 1569 1584 1595 1605 1624 1641 1660 1675 1697 1711 1727 1741 1758 1769
>>>>> 1788 1807 1822 1841 1858 1880 1899 1909 1923 1935 1957 1971 1985 1997
>>>>> 2014 2033 2044 2056 2076 2095 2110 2124 2141 2162 2181 2195 2210 2227
>>>>> 2243 2265 2275 2294 2316 2331 2350 2357 2369 2386 2402 2416 2430 2449
>>>>> 2473 2495 2512 2519 2531 2543 2558 2569 2576 2595 2605 2620 2641 2660
>>>>> 2680 2692 2714 2731 2745 2764 2774 2788 2799 2816 2832 2856 2873 2887
>>>>> 2908 2925 2942 2954 2965 2980 2990 3000 3019 3033 3057 3074 3093 3107
>>>>> 3126 3141 3160 3181 3191 3202 3223 3239 3260 3279 3290 3307 3318 3339
>>>>> 3360 3380 3392 3416 3428 3440 3456 3466 3485 3507 3521 3541 3551 3573
>>>>> 3594 3614 3633 3650 3667 3678 3695 3710 3725 3749 3764 3781 3791 3806
>>>>> 3828 3847 3864 3886 3905 3922 3936 3953 3977 3984 3991 4015 4034 4054
>>>>> 4073 4090 4102 4121 4138 4162 4174 4186 4197 4209 4221 4245 4260 4271
>>>>> 4278 4297 4311 4321 4338 4353 4364 4374 4393 4410 4429 4444 4466 4480
>>>>> 4496 4510 4527 4538 4557 4576 4591 4610 4627 4649 4668 4678 4692 4704
>>>>> 4726 4740 4754 4766 4783 4802 4813 4825 4845 4864 4879 4893 4910 4931
>>>>> 4950 4964 4979 4996 5012 5034 5044 5063 5085 5100 5119 5126 5138 5155
>>>>> 5171 5185 5199 5218 5242 5264 5281 5288 5300 5312 5327 5338 5345 5364
>>>>> 5374 5389 5410 5429 5449 5461 5483 5500 5514 5533 5543 5557 5568 5585
>>>>> 5601 5625 5642 5656 5677 5694 5711 5723 5734 5749 5759 5769 5788 5802
>>>>> 5826 5843 5862 5876 5895 5910 5929 5950 5960 5971 5992 6008 6029 6048
>>>>> 6059 6076 6087 6108 6129 6149 6161 6185 6197 6209 6225 6235 6254 6276
>>>>> 6290 6310 6320 6342 6363 6383 6402 6419 6436 6447 6464 6479 6494 6518
>>>>> 6533 6550 6560 6575 6597 6616 6633 6655 6674 6691 6705 6722 6746 6753
>>>>> 6760 6784 6803 6823 6842 6859 6871 6890 6907 6931 6943 6955 6966 6978
>>>>> 6990 7014 7029 7040 7047 7066 7080 7090 7107 7122 7133 7143 7162 7179
>>>>> 7198 7213 7235 7249 7265 7279 7296 7307 7326 7345 7360 7379 7396 7418
>>>>> 7437 7447 7461 7473 7495 7509 7523 7535 7552 7571 7582 7594 7614 7633
>>>>> 7648 7662 7679 7700 7719 7733 7748 7765 7781 7803 7813 7832 7854 7869
>>>>> 7888 7895 7907 7924 7940 7954 7968 7987 8011 8033 8050 8057 8069 8081
>>>>> 8096 8107 8114 8133 8143 8158 8179 8198 8218 8230 8252 8269 8283 8302
>>>>> }
>>>>> }
>>>>> axis ( 0.0, 0.0, 1.0 )
>>>>> }
>>>>> }
>>>>>
>>>>> harmonic {
>>>>> colvars ferritin
>>>>> centers 130.0
>>>>> forceConstant 0.2
>>>>> # targetCenters 134.0
>>>>> # targetNumSteps 1
>>>>> }
>>>>>
>>>>>
>>>>> Olga
>>>>>
>>>>
>>>> --
>>>> Amy Rice
>>>> Ph.D. Student
>>>> Physics Department
>>>> Illinois Institute of Technology
>>>
>

This archive was generated by hypermail 2.1.6 : Sun Dec 31 2017 - 23:20:41 CST