Re: AMBER: question about formatted mdrcrd files...

From: Cenk Andac <cenk_andac.yahoo.com>
Date: Sun, 22 Jul 2007 17:51:02 -0700 (PDT)

Hi Prof. Case,

Thanks for replying.

I use sander (version 9) and I compiled it with intel fortran x86_64. You are right, I initially had "****.*******" in places of 13 digits in my mdcrd and I placed the missing coordinated manually in the mdcrd file by looking at the coordinates in the last snapshot of the corresponding trajectory.

Well, I have tried reducing the 13 digits to 12 digits manually in the mdrcd file. However, the next sander job gave a segmentation fault error message when I used the modified mdrcd file as an input. I believe, this is because the decimal points for the newly added numbers do not nicely line up. For instance, it is impossible to line up decimal points for the first coordinates in Lines 1 and 2 below.

-1079.930054
-286.6010762

The problem is that the first coordinate in Line 1 has 5 digits (including "-") on the left side of the decimal point. As far as I know, in a formatted mdcrd, only 4 digits are allowed on the left side of a decimal point.

Is there any way to make sander accept coordinates with 5 digits on the left side of coordinates without ever disturbing positions of nicely line-up decimal points in an mdrcd file?

thanks for your in advance..

regards,

jenk.




"David A. Case" <case.scripps.edu> wrote: On Sun, Jul 22, 2007, Cenk Andac wrote:
>
> I have run a 110 ns MD for an RNA system explicitely solvated by TIP3 water
> molecules. I would like to run another 50 ns MD for the system before
> I start analyzing the trajectory. However, there appears to be a problem
> with the last mdcrd file (acquired at the end of 110 ns MD). After going
> through each line in the mdcrd file i figured that the format of the mdcrd
> file was disrupted by huge numbers as examplified below (talking about the
> first line compared to the conventional second line):
>
> Line 1) -1079.9300543 -31.8968786 259.2745548-1079.0899662 -31.8930939
> 258.0129848
> Line 2) -286.6010762 12.6202845 20.8281542-286.9375394 13.5078975
> 20.704998
>
> Is there any way to correct (??reformat??) line 1 (especially the first number in Line 1) so that I can continue my MD experiment?
>

If there are only a handful of such entries, you could reformat by hand, to
remove the 13-digit entries (so all entries are 12 digits). Otherwise, you
need to go back to the last restart before this one, and start there, setting
iwrap=1.

[It would be on interest to know what compiler you used: all compilers I am
aware of would print "***" in a field if 13 characters were needed by only
12 were called for in the format statement (which in this case is 6f12.7).
But your system seems more like "C", in that it added an extra digit and
pushed everything else to the right. And, is this sander or pmemd? Which
version?]

...regards...dac



       
---------------------------------
Get the Yahoo! toolbar and be alerted to new email wherever you're surfing.
-----------------------------------------------------------------------
The AMBER Mail Reflector
To post, send mail to amber.scripps.edu
To unsubscribe, send "unsubscribe amber" to majordomo.scripps.edu
Received on Wed Jul 25 2007 - 06:07:10 PDT
Custom Search