I agree with Dan, you can use whatever measure you want for "native", and
as long as you understand what structures you're including, it's
technically valid. The difficulty comes when you try to compare to
experiments, especially if you're using the melting curve to validate your
work. You should probably use metrics that are closely related to the
experimental observables. Be very careful not to compare things that don't
probe the same structure regions, or even the same kinds of "structure".
For example, using RMSD to compare to experimental CD data could be
problematic, unless you know that the two metrics are highly correlated. In
other words, the problem isn't in how you calculate the melting curve, it's
in how you interpret it.
for some recent examples from my lab, look at
1. Improved Generalized Born Solvent Model Parameters for Protein
Simulations
Hai Nguyen, Daniel R. Roe, and Carlos Simmerling
Journal of Chemical Theory and Computation *2013* *9* (4), 2020-2034
http://pubs.acs.org/doi/abs/10.1021/ct3010485
On Mon, Oct 7, 2013 at 10:39 AM, Daniel Roe <daniel.r.roe.gmail.com> wrote:
> Hi,
>
> On Sun, Oct 6, 2013 at 9:59 PM, Gargi Borgohai <gargib2011.gmail.com>
> wrote:
> > Will you please confirm me whether I can use the population of native
> structure
> > obtained directly from cluster structure analysis for generation of
> > melting curves or not.
>
> You *can* use cluster population, but as you have found it can be
> problematic depending on how well-defined your clusters are. It's not
> clear what "high" and "low" RMSD are for you, but if your "native"
> cluster (I assume you're assigning this based on similarity of the
> cluster representatives to a well-defined native structure) contains
> non-native structures you may need to adjust epsilon in addition to
> the number of clusters.
>
> Keep in mind as well there are other ways to define what is "native"
> (e.g. you could use native contacts or hbonds), and each may result in
> different looking melting curves. Depending on what your ultimate goal
> is, other metrics may be appropriate. For example, if you're comparing
> to a melting curve obtained via CD, using a native metric based on
> secondary structure content may be more appropriate.
>
> Hope this helps,
>
> -Dan
>
> --
> -------------------------
> Daniel R. Roe, PhD
> Department of Medicinal Chemistry
> University of Utah
> 30 South 2000 East, Room 201
> Salt Lake City, UT 84112-5820
> http://home.chpc.utah.edu/~cheatham/
> (801) 587-9652
> (801) 585-6208 (Fax)
>
> _______________________________________________
> AMBER mailing list
> AMBER.ambermd.org
> http://lists.ambermd.org/mailman/listinfo/amber
>
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Mon Oct 07 2013 - 08:00:06 PDT
