[buster-discuss] abasic refinement.
Francis Reyes
Francis.Reyes at Colorado.EDU
Thu Jun 12 13:01:32 CEST 2014
Actually Ed's suggestion was exactly what I was looking for.
Setting the base occupancies to zero (but sugars and phosphates to 1) gave me "pseudo abasic" refinement.
It was probably much better than completely removing the atoms from the input PDB in that the atoms could be included for geometric reasons (similar to riding hydrogens) but not included in the calculation of scattering F.
Also zero occupancy atoms are included in buster's solvent and non uniform prior calculation (correct me if I'm wrong Claus or Clemens), as judged by specific buster options excluding these atoms.
The issue was that by completely removing them from the PDB, buster was comparing it to its own definition of what the base should be (from nuclgeo.dat) . It realized that more than half of the atoms were missing, and instead got lost as to which restraints to use. Conceivably I could have made my own restraint file, but considering how buster treats zero occ atoms (as explained above), it was better to include them as "riding".
Thanks all,
F
> On Jun 12, 2014, at 4:39 AM, FOOS Nicolas <nicolas.foos at synchrotron-soleil.fr> wrote:
>
> Just to be sure of what Buster did,
> Did it complain about something at the begining of the refinement ?
> Maybe you can check if the problem come from missing atoms, or if it come from the misundestanding DNA.
> You can also generate your own restraint for your molecules and give this resraint to Buster.
>
> Hope to help.
> ________________________________________
> De : buster-discuss [buster-discuss-bounces at globalphasing.com] de la part de Ed Pozharski [ccp4pozharski at gmail.com]
> Envoyé : mercredi 11 juin 2014 20:04
> À : Francis Reyes; buster-discuss at globalphasing.com
> Objet : Re: [buster-discuss] abasic refinement.
>
> Ok, how about polyA with side chain occupancies set to zero?
>
>
> Sent on a Sprint Samsung Galaxy S® III
>
>
> -------- Original message --------
> From: Francis Reyes
> Date:06/11/2014 1:59 PM (GMT-05:00)
> To: Ed Pozharski , buster-discuss at globalphasing.com
> Subject: Re: [buster-discuss] abasic refinement.
>
> Ed,
>
> The problem is not in making the PDB. I've made it.
>
> For example,
>
> ATOM 1216 P DG C 9 24.074 -29.421 24.416 1.00 20.00 P
> ATOM 1217 OP1 DG C 9 22.922 -29.968 23.666 1.00 20.00 O
> ATOM 1218 OP2 DG C 9 23.858 -28.739 25.711 1.00 20.00 O
> ATOM 1219 O5' DG C 9 25.147 -30.586 24.645 1.00 20.00 O
> ATOM 1220 O3' DG C 9 28.318 -32.882 25.868 1.00 20.00 O
> ATOM 1221 C1' DG C 9 28.671 -30.101 25.829 1.00 20.00 C
> ATOM 1222 C2' DG C 9 27.497 -30.719 26.571 1.00 20.00 C
> ATOM 1223 C3' DG C 9 27.286 -31.929 25.672 1.00 20.00 C
> ATOM 1224 C4' DG C 9 27.428 -31.312 24.285 1.00 20.00 C
> ATOM 1225 C5' DG C 9 26.129 -30.850 23.651 1.00 20.00 C
> ATOM 1226 O4' DG C 9 28.312 -30.175 24.464 1.00 20.00 O
> and no base atoms.
>
> BUSTER just has issues with obtaining the restraints for it and tries to automatically generate its own restraints .
>
> Is it possible to increase the number of atoms missing before triggering automatic restraint generation?
>
> I tried a combination of NeverGenerateDictionary="yes" but it gets shuttle to PDB2TNT which doesn't know about the connectivity between the residues. (it treats each base as a monomer and will break the interresidue bond in refinement).
>
>
>
>
> F
>
>> On Jun 11, 2014, at 1:53 PM, Ed Pozharski <ccp4pozharski at gmail.com> wrote:
>>
>> You can make polyA and then delete "side chains" (in coot or with a grep one liner). This way it will also be easier to build them back in (although you may need format conversion post-buster to get coot to work and I vaguely remember coot having some problems with truncation).
>>
>>
>> Sent on a Sprint Samsung Galaxy S® III
>>
>>
>> -------- Original message --------
>> From: Francis Reyes
>> Date:06/11/2014 1:39 PM (GMT-05:00)
>> To: buster-discuss at globalphasing.com
>> Subject: [buster-discuss] abasic refinement.
>>
>> Hi all,
>>
>> I'm refining a protein:dna complex, and I'm yet uncertain to the sequence identity of the bases. Instead I'd like to to just simply place the phosphate backbone of the double stranded DNA and removed the bases (abasic DNA).
>>
>> I receive the following error:
>>
>>>>> ab_pdb2tnt >>>
>>>>> ab_pdb2tnt >>> Extracted from file 06/01-BUSTER/Rigid-1/ab_pdb2tnt.log :
>>>>> ab_pdb2tnt >>>
>>>>> ab_pdb2tnt >>> WARNING : 10 atoms (out of 11) for residue "C" in PDB file
>>>>> ab_pdb2tnt >>> have names that are not present in the suggested
>>>>> ab_pdb2tnt >>> dictionary
>>>>> ab_pdb2tnt >>> /groups/gonen/home/reyesf/xtal/buster/tnt/data/nuclgeo.dat.
>>>>> ab_pdb2tnt >>> Therefore we will force the generation of a new
>>>>> ab_pdb2tnt >>> dictionary based on the current coordinates! For
>>>>> ab_pdb2tnt >>> details see below:
>>
>>
>> and then restraint generation by grade ends with error.
>>
>>
>> Is it possible to refine abasic DNA?
>>
>> Thanks,
>>
>> F
>>
>> _______________________________________________
>> buster-discuss mailing list
>> buster-discuss at globalphasing.com
>> https://www.globalphasing.com/mailman/listinfo/buster-discuss
>
More information about the buster-discuss
mailing list