Thursday, 22 September 2011

refmac and DNA

From: Ed Pozharski
Date: 8 September 2011 14:35

After switching (finally) to 6.2.0 and therefore to Refmac 5.6.0117 I
have found a problem working with DNA that I have not seen with
6.1.13/5.5.0109.  Namely,

- if I use the pdb file produced by Coot (0.7.pre-1.3470) that seems to
output DNA as Ad/Td/Gd/Cd no matter what the input names were, refmac
fails with the warning that it found a new monomer.  It appears that it
stumbles upon the very first thymidine, but in a strange twist it
reports the problematic residue having the name "DY"!

- if I use the pdb file previously produced by refmac, which has the
A/T/G/C as residue names, it fails too but now complains about the "new"
monomer named "T".

- the workaround I found is to rename all the thymidines to "DT".  It is
a bit annoying since coot keeps renaming them (well, not refmac/ccp4
problem per se) and I have to rename back (easily scripted task, of
course).  What is peculiar is that Ad/Gd/Cd don't need to be renamed
(does this have anything to do with thymidine being the only one that
changes residue name in RNA?).

Has anyone else seen this or it's something specific to my setup?

Cheers,

Ed.

--
After much deep and profound brain things inside my head,
I have decided to thank you for bringing peace to our home.
                                   Julian, King of Lemurs

----------
From: Miguel Ortiz Lombardía
Weird. I'm using Coot 0.7-pre1.3628 ( a bit ahead of yours but not so
much? or maybe yes? ) and DNA residues are saved in what I think it is
PDB v3 format (  DA, DC, DG, DT ) These files are properly understood by
refmac 5.6.0119, which at the end produces coordinate files in the same
format ( so DNA residues are:  DA, DC, DG, DT )

The only heck with coot/thymidines is that if you "add a terminal" T or
mutate another DNA residue to T, it lacks its C7 ( formerly C5M ) atom.
However, if you build an ideal T, it has C7... My workaround is to add
those pesky C7 atoms in dummy positions and refine.
Not what I find (see above)
Perhaps the Nd/DN issue was solved between revisions 3470 and 3628 ?

Best,


--
Miguel


----------
From: Ed Pozharski 

Indeed.  I just built the rev 3633 and it works fine.  Autobuild scripts
used to have some problems on Lucid, but this time it worked perfectly.

This evolved into a coot question, but I am still perplexed as to why
refmac would complain only about T and not A/G/C.

----------
From: Miguel Ortiz Lombardía
A, C and G are RNA nucleotides. T is (mostly) not, its RNA-equivalent is
uridine phosphate, U.

Cheers,

----------
From: Ed Pozharski
Right, that was my suspicion.  But I thought that RNA bases would be Xr,
not Xd.  Plus, refmac does not complain about missing oxygens.  Anyway,
the latest coot rocks, as usual, now if I only can figure out who is
Linda...

Cheers,

Ed.



--
Coot verendus est

----------
From: Miguel Ortiz Lombardia
Le 11/09/2011 00:23, Ed Pozharski a écrit :
My understanding is that the Xr/Xd scheme has been dropped from the
latest refmac5/coot to comply with the X/DX (and OP1, OP2, prime signs
instead of asterisks...) standard from the PDB/NDB.


No comments:

Post a Comment