VMD-L Mailing List

From: t.shivam_at_iitg.ernet.in
Date: Sun Aug 06 2017 - 23:07:36 CDT

> On Sun, Aug 6, 2017 at 10:30 PM, <t.shivam_at_iitg.ernet.in> wrote:
>
>>
>>
>> > On Sun, Aug 6, 2017 at 10:51 AM, <t.shivam_at_iitg.ernet.in> wrote:
>> >
>> >> Dear vmd users,
>> >>
>> >> I have a protein solvated in a water box whose net charge is
>> >> 7.999999954365194, so when i am using autoionize to neutralize my
>> system
>> >> it is showing a charge of -4.563480615615845e-8, i also tried to add
>> 8
>> >> chlorine ions instead of going for neutralize option (anyway its more
>> or
>> >> less the same thing), so please kindly help me with this problem.
>> >>
>> >
>> > ​there is no problem with autoionize. the problem is between your
>> chair
>> > and
>> > your computer.​ you are doing floating point math here​​, so
>> -4.5e-8
>> > is
>> > essentially zero and most certainly zero when rounded to the precision
>> at
>> > which partial charges are specified in the CHARMM force field.
>> >
>> > axel.
>>
>>
>> this might sound funny to you but previously i have solvated my protein
>> in
>> a smaller box but the charge of the protein is same
>> (7.999999954365194)so
>> at that when i neutralized my system, vmd was showing a perfect
>> zero(0.0),
>>
>
> ​this is not funny. it only shows, that ​you seem unaware of the
> idiosyncrasies of doing math with floating point numbers.
>
>
>> so i saved the files of the ionized system, but after that i also
>> created
>> a bigger box, and now when i am neutralizing it, it is showing
>> -4.563480615615845e-8, and the strangest part is that when i am checking
>> my previous system (smaller box showing charge 0.0) it is also showing
>> -4.563480615615845e-8,why is that happening?i agree that the value
>> -4.563480615615845e-8 is almost zero.
>>
>
> ​this is happening, because you are doing floating point math (and are
> outputting results with a precision that exceeds the precision of
> storage).
> VMD stores charges as float, and quite a few numbers cannot be represented
> exactly with floating point representations.​
>
> i suggest you study the following two blog posts:
> http://blog.reverberate.org/2014/09/what-every-computer-programmer-should.html
> http://blog.reverberate.org/2016/02/06/floating-point-demystified-part2.html
>
> p
> ​lease keep in mind, the origin of the non-exactly-zero is the summing
> of
> the partial charges which results in not-exactly-eight. just what do you
> get, when you subtract 8 from 7.999999954365194 ?? there is no mystery at
> all here.
>
> axel.
>
>
>> regards
>> shivam
>> >
>> >
>> >>
>> >> regards
>> >> shivam
>> >>
>> >>
>> >
>> >
>> > --
>> > Dr. Axel Kohlmeyer akohlmey_at_gmail.com http://goo.gl/1wk0
>> > College of Science & Technology, Temple University, Philadelphia PA,
>> USA
>> > International Centre for Theoretical Physics, Trieste. Italy.
>> >
>>
>>
>
>
> --
> Dr. Axel Kohlmeyer akohlmey_at_gmail.com http://goo.gl/1wk0
> College of Science & Technology, Temple University, Philadelphia PA, USA
> International Centre for Theoretical Physics, Trieste. Italy.
>

yes sir i was unaware of it, i am quite novice to the computational world,
thanks for your suggestions and the link, i have one follow up question
regarding the same, since i have obtained the psf of my protein using opls
topology, but autoionize is using charmm topology to generate the psf
files will it create problem in my simulations since i wish to use oplsAA
force field for my simulations,if yes how can i use opls topology to add
ions to my system.

regards
shivam