[buster-discuss] AdjustModifiedNucleotides

Edwin Pozharski pozharskibb at gmail.com
Wed Aug 22 00:50:49 CEST 2018


Thanks - I suspected that.  There is an option - SequenceFileGeneration -
that allows to use pdb2seq instaeaf MakeLINK, but when I do that refine run
fails while spitting out a lot of errors about duplicate atoms.

Your suggested method almost works, except that pdb2seq does not recognize
DA/DT/DG/DC as nucleotides and leaves them without a type (MakeLINK labels
them as dSUGPHOS).  I recall there used to be an issue where I had to use
A/T/G/C naming to get proper links, perhaps that was back when pdb2seq was
teh default.  Another thing is that pdb2seq still doesn't assign dSUGPHOS
to non-standard nucleotide - instead, I get the following comment next to
it: "this might be a modified DNA base (consider modifying this sequence
file if it is part of a chain)".  Does not seem like the modified
nucleotide is "adjusted" by pdb2seq either.

So here is the method that works without LINK records:

1. Use MakeLINK to generate sequence file like so
MakeLINK -p model.pdb
which creates model.pdb.seq
2. Edit model.pdb.seq to make sure that (i) preceding residue is listed as
"... DX <N> dSUGPHOS" where N is the residue number for the non-standard
nucleotide - it defaults to "... DX NULA d3'END"; (ii) non-standard
nucleotide is listed as "...XXX <N+1> dSUGPHOS"; (iii) following residue is
listed as "...DX <N+2> dSUGPHOS".  If I run into this issue too often, it's
easy to spin out a simple python script to parse and adjust the sequence
file.
3. Add "-Seq model.pdb.seq" to refine command.

This works and has the additional benefit in my specific case -
non-standard nucleotide is in tow conformations and this way there is no
need to define a pair of LINK records.  Of course, the shortcoming here is
that sequence file would need to be regenerated if I add anything but
waters to the mix.

Thanks for prompt response!

Ed.



On Tue, Aug 21, 2018 at 5:51 PM ClAuS Flensburg <claus at globalphasing.com>
wrote:

> Hi Ed,
>
> On Tue, Aug 21, 2018 at 05:27:08PM -0400, Edwin Pozharski wrote:
> > Hello,
> >
> > could someone confirm that AdjustModifiedNucleotides option actually
> > works?
>
> it does ... for the helper program pdb2seq. However, the default
> method to generate the sequence file within BUSTER is to use the
> alternative helper: MakeLINK.
>
> > IIUC, it is supposed to tell tnt to treat residues that contain
> > proper sugar-phosphate atom names as DNA/RNA.  I have some modified
> > nucleotides and always get Gelly sanity check error telling me that there
> > is bad contact between its P/O3' and preceding/following nucleotide
> O3'/P.
> > The issue can be resolved by adding proper LINK records to the input PDB
> > file, but shouldn't this be unnecessary given
> > that AdjustModifiedNucleotides defaults to "yes"?
>
> as an alternative to adding the LINK records you should be able to
> generate the sequence file using pdb2seq first eg.
>
>     % pdb2seq -p model.pdb -o model.seq
>     % refine -m data.mtz -p model.pdb -Seq model.seq ...
>
> but, yes, better would be to have MakeLINK take
> AdjustModifiedNucleotides into account too.
>
>
> Regards,
>
> ClAuS (for BUSTER Develpers)
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.globalphasing.com/pipermail/buster-discuss/attachments/20180821/eb911956/attachment.html>


More information about the buster-discuss mailing list