Re: [AMBER] Ptraj Hbond query.

From: Thomas Cheatham <tec3.utah.edu>
Date: Tue, 10 May 2011 20:49:05 -0600 (Mountain Daylight Time)

> This question my seem somewhat trivial, however...here goes. From what
> I gather, the latest version of the ptraj H-bond analysis tool outputs a
> string of 50 characters, each indicating the H-bond occupancy for each
> 50th of the trajectory. For my job, I use the basic skeleton of the
...
> The problem I have is that in my output I'm only getting 48
> symbols....indicating each symbol only accounts for 1/48th of the
> simulation. Has this been seen before and is there any explanation for
> this?

Let's just say that this simple visual output is only intended to be
qualitative to provide a simple text based visual of what h-bonds are
fully formed versus partially formed... I never intended this to be more
than a qualitative visual aid and I could have messed up on exact
intervals, etc for this simple plot. I did this to avoid having to
actually plot each h-bond...

Normally I only need qualitative insight to decide what to go digging for
with more quantitative / accurate information in subsequent analyses.
The simple text based output usually points me to important interactions.
However, if people are "mining" this qualitative plot, I could look to
check indices more accurately, i.e. 48/50. To understand the printing,
see the code for transformHBond in actions.c...

If you or others have ideas regarding what you really want to see, this
could be useful information. Previously we only printed the hbond
information in 10 characters; 50 (or 48) is better, but, we need to
understand what people really want. What could ptraj output that would
facilitate interpretation and analyses? I have seen from posts from the
reflector that this is still confusing, i.e. a darker/more full character
means fully occupied (.) while a less dense character (.) or space means
less occupied. Perhaps we need to improve the manual.

I cannot promise that I or others will implement good ideas -- as always
with AMBER we are welcome to others updating the code at anytime as this
is a community effort and usually we'll help guide as much as we can!!! --
but feedback is useful.

--tec3


_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Tue May 10 2011 - 20:00:02 PDT
Custom Search