AMBER Archive (2008)

Subject: Re: AMBER: silhouette width/coefficient after clustering by PTRAJ9.9

From: Dong Xu (d1xu_at_ucsd.edu)
Date: Mon May 12 2008 - 23:09:11 CDT

Hi Jianyin and sychen,

I'm also very interesting in using ptraj to perform clustering
analysis. However, I wasn't able to locate any documentation on this
topic, not even on ptraj official web site:
http://www.chpc.utah.edu/~cheatham/ptraj.html

Maybe I have missed something here, would you please email some
instructions as well as the scripts that calculate the silhouette
width?

Many thanks!

Dong Xu

On Fri, May 9, 2008 at 1:05 PM, Jianyin Shao <jyshao2004_at_gmail.com> wrote:
> Hi, sychen,
>
> I think we did replace the mask parser with a new one. I didn't realize this
> discrepancy.
> I would suggest you to use "@461-474,494-507" under the new version. If you
> insist using the old parser, you can replace the call of function
> processAtomMaskDetailedVH() back to processAtomMaskDetailed() in lines 1352
> and 1359 in ptraj.c file.
>
> Please send me a copy of your failed SOM clustering script.
>
> Best,
>
> Jianyin
>
>
> On 5/8/08, sychen <u8613020_at_msg.ndhu.edu.tw> wrote:
> > Hi, Jianyin Shao,
> > Thank you for your kind help. It works fine.
> > However I got 2 problems in PTRAJ,
> > 1. Does atom selection differ in the latest version?
> > \-/ Version: "AMBER 9.0 integrated" (April 2006)
> > -/- Executable is: "ptraj"
> > /-\
> > \-/ Residue labels:
> >
> > (omit)
> >
> > PTRAJ: matrix mwcovar name a16a17.covar @461-_at_474,@494-_at_507 mass
> > Mask [@461-_at_474,@494-_at_507] represents 28 atoms
> >
> >
> > \-/
> > -/- PTRAJ: a utility for processing trajectory files
> > /-\
> > \-/ Version: "Cheatham lab development full version 9.9g" (1 April
> 2008)
> > -/- Executable is: "ptraj"
> > /-\
> > \-/ Residue labels:
> >
> > (omit)
> >
> > PTRAJ: matrix mwcovar name a16a17.covar @461-_at_474,@494-_at_507 mass
> > Mask [@461-_at_474,@494-_at_507] represents 4 atoms
> >
> > I also trid the latest version 9.9h, and the result was as same as 9.9g
> gave.
> >
> > 2. Does there any paticular environment to be modify before compiling
> PTRAJ 9.9g(h)?
> > The clustering always fail as SOM algorithm was used and the error
> message was as follows,
> >
> > The first pass through the trajectory is complete.
> > Clustering using the SOM algorithm.
> > --------------------------------------------------------------------
> > Error in file cluster.c line 6520 when calling SafeMalloc() to alloc of
> -547749984 bytes.
> >
> >
> >
> > Sincerely,
> > sychen
> >
> >
> > On Fri, 2 May 2008 17:27:44 -0600
> > "Jianyin Shao" <jyshao2004_at_gmail.com> wrote:
> >
> > > Hi, sychen,
> > >
> > > Attached is a perl script to calculate the silhouette width for the
> ptraj clustering output.
> > > Please run "silhouette.pl -h" to see the usage. If you have any
> questions or suggestions, please let me know. If I incorporate the code into
> the ptraj, I will let you know.
> > >
> > > Thanks,
> > >
> > > JS
> > >
> > > On 5/2/08, sychen <u8613020_at_msg.ndhu.edu.tw> wrote: Dear all,
> >
> >
> >
> > > I need a program or script to calculate silhouette
> width/coefficient
> > > after clustering by PTRAJ9.9, because I'd like to make for sure
> of the
> > > quality of clustering numbers/algorithms and I could not find
> such
> > > function included in PTRAJ so far.
> > > Thank you very much!
> > >
> > >
> > >
> > > Best regards,
> > > sychen.
> > >
> > >
> -----------------------------------------------------------------------
> > > The AMBER Mail Reflector
> > > To post, send mail to amber_at_scripps.edu
> > > To unsubscribe, send "unsubscribe amber" to majordomo_at_scripps.edu
> > >
> > >
> >
> >
> > --
> > sychen <u8613020_at_mail.ndhu.edu.tw>
> >
> >
>
>
-----------------------------------------------------------------------
The AMBER Mail Reflector
To post, send mail to amber_at_scripps.edu
To unsubscribe, send "unsubscribe amber" (in the *body* of the email)
      to majordomo_at_scripps.edu