[RNABuilder-help] Re: chains and rnabuilder

[Samuel Flores]

On Jan 7, 2011, at 3:28 PM, Pete Kekenes-Huskey wrote:

> Hi Sam,
> Thanks for getting back to me. I will gladly discuss this via the
> rnatoolbox-help list? Should I start a new thread?

Just please send it to the list as I am doing here.

> 
> In the meantime, could you point me to the PSB talk you mentioned?
> 
> I've verified that I am using chain Q for my template (based on a
> crystal structure) and C for my input structure, as you did in the
> RNABuilder documentation.

I think you mean that you used C for your model, or threaded chain.  You have no chain C in your input file.

> What seems to happen is that the first chain
> (residues 1-9) seem to assume the template conformation, while the
> remaining residues (10-12) of the second chain appear as a random
> coil.
> 

Your first problem is that all your O2' atoms are in the wrong position.  Easiest thing to do is delete all these atoms, and let RNABuilder guess their positions.  

However your main problem is that you have a gap between residues 9 and 10.  The "TER" record suggests you think of these as separate chains, and yet you are using the same chain ID for residues 1-9 and 10-12, which suggests the opposite.  I am going to go with the second interpretation.  The solution is the following.

I pride myself in supporting your ability to read in some rather unusual and chemically improbable structures.  You can leave out any atom you want and let RNABuilder guess its position, with the caveat that any atoms which are children of that atom on the multibody tree will appear in an uncontrolled position.  This is no problem for side chains, but in this case you are missing a main chain atom -- the P atom of residue 10.  I suggest you simply rename "H5T" of residue 10 to "P".  It won't be chemically correct but it will allow you to match the input structure the way you intend.


> last.1.pdb is my template and the threading.dat file essentially is
> using the same sequence for as the template.
> 
> THanks again for your help!
> 
> pete
> 
> 
> On Tue, Jan 4, 2011 at 4:22 PM, Samuel Flores <samuelfloresc at gmail.com> wrote:
>> Hi Pete,
>> Thanks for your interest in RNABuilder.Threading is one of the most popular
>> applications of rnabuilder.  You may be interested in reading my PSB 2011
>> paper on multiple-template homology modeling.  I can provide the input files
>> for that project.
>> Can we use the rnatoolbox-help list for this, or is it confidential?
>>  thanks.
>> As for your question, please make sure your chain ID's for the two chains
>> are different and that they match in the rnabuilder input file vs. structure
>> file.
>> If this is not the issue please send me your last.n.pdb and the rnabuilder
>> input file, (e.g. contacts.dat).
>> Sam
>> You have been contacted by a member of Simtk.org.
>> 
>> From: Pete Kekenes-Huskey
>> Email: huskeypm at wag.caltech.edu
>> 
>> The e-mail was sent via the "contact" link on Simtk.org by the above member.
>> It is not screened by Simtk.org or its staff.
>> 
>> 
>> *** MESSAGE BEGINS ***
>> 
>> Hi Sam
>> I've been using your RNABuilder program to thread short RNA sequences. I
>> have a target molecule which consists of two chains separated by a TER card.
>> It appears that when the 'last.n.pdb' structure (target) is loaded, the
>> first chain preserves the original coordinates, but the second chain assumes
>> a random coil form and is adjoined to the first chain. Do you have any
>> suggestions for forcing the second chain to preserve its original
>> coordinates?
>> 
>> I can provide input files if you email me at huskeypm at gmail.com
>> Thanks for your help!
>> 
>> Best,
>> pete
>> 
> 
> 
> 
> -- 
> ======================
> Peter Kekenes-Huskey, Ph.D.
> Postdoctoral Researcher
> UCSD Dept of Pharmacology
> 4205 Urey Hall
> 9500 Gilman Avenue
> La Jolla, CA 92093
> ======================
> <threading.dat><last.1.pdb>