Previous Thread
Next Thread
Print Thread
CHARMM MC energy minimization random seed question
#38317 08/06/21 01:48 AM
Joined: Aug 2021
Posts: 1
Y
yycho Offline OP
Forum Member
OP Offline
Forum Member
Y
Joined: Aug 2021
Posts: 1
Hi everyone

I am developing a protein affinity evaluation program using the CHARMM MC module to minimize protein mutation energy.

I have a problem with reproducibility.
I think that it is a CHARMM MC random seed problem. I got slightly different results for each other computing servers.
I tried to give MC ISEEd to fix a random seed number, but it still is not fixed.

Here is my CHARMM MC code
""


!MC script
!Print heavy atoms with unknown coordinates
coor print sele ( .not. INIT ) .and. ( .not. hydrogen ) end

define ROT sele .byres. (.not. INIT .and. (type CB .or. type CG*)) .and. .not. resn *PRO end
define XXX sele ROT .and. type CA end
!Print heavy atoms with unknown coordinates
coor print sele ( .not. INIT ) .and. ( .not. hydrogen ) end

define ROT sele .byres. (.not. INIT .and. (type CB .or. type CG*)) .and. .not. resn *PRO end
define XXX sele ROT .and. type CA end
set nres = ?nsel
set ires = 1
label build_rot
define YYY sele .byres. ( XXX .subset. @ires ) end
define ZZZ sele .bonded. YYY .and. .not. YYY .and. .not. (type N .or. type C) end
if ?nsel .gt. 0 define ROT sele ROT .and. .not. YYY end
incr ires by 1
if ires .le. @nres goto build_rot

ic para
ic build
hbuild


ENERGY

define ROT sele .byres. ( resid @mutresn .and. (type CB .or. type CG*)) .and. .not. resn *PRO show end
define XXX sele ROT .and. type CB end

MC ISEEd 3355
if ?nsel .gt. 0 then
cons fix sele .not. ROT end
MOVE ADD MVTP TORS DMAX 360.0 sele ( ROT .and. type CA ) .and. segid PRO0 end -
sele ( ROT .and. type CB ) .and. segid PRO0 end
MOVE ADD MVTP TORS DMAX 360.0 sele ( ROT .and. type CA ) .and. segid P2PRO0 end -
sele ( ROT .and. type CB ) .and. segid P2PRO0 end

MC NSTEP 5000
MOVE DELETE MVIN 1
cons fix sele none end
endif

""

CHARMM MC Manual
ISEEd The seed for the random number generator. If it is not
specified, it is unchanged, so that a script can be seeded
once initially and then loop over an MC command and yield
different behavior with each call."




CHARMM MC manual indicates that we can only fix the initial random number with ISEEd?
Is there any other option that CHARMM MC must give for fixing the results?




Thank you

Re: CHARMM MC energy minimization random seed question
yycho #38318 08/06/21 03:57 PM
Joined: Sep 2003
Posts: 8,579
Likes: 11
rmv Offline
Forum Member
Offline
Forum Member
Joined: Sep 2003
Posts: 8,579
Likes: 11
I suggest reading the "random" info file, and trying a few short tests based on the information in that file. It is entirely possible that the MC info file has not been kept up to date by the author.


Rick Venable
computational chemist


Moderated by  BRBrooks, lennart, rmv 

Link Copied to Clipboard
Powered by UBB.threads™ PHP Forum Software 7.7.4
(Release build 20200307)
Responsive Width:

PHP: 7.3.31-1~deb10u1 Page Time: 0.008s Queries: 18 (0.005s) Memory: 0.7201 MB (Peak: 0.7574 MB) Data Comp: Off Server Time: 2021-11-28 11:17:32 UTC
Valid HTML 5 and Valid CSS