Hi Paul, I would have them both in PDB file with different non-blanc altLocs and arbitrary starting occupancies and that will work in refinement (in phenix.refine for sure, can't tell for other programs). Pavel
On Fri, Aug 23, 2013 at 8:39 AM, Paul Paukstelis <shocksofmig...@gmail.com>wrote: > Greetings, > > We have been working on a few DNA crystals in which the asymmetric unit > contains a stoichiometric (or nearly so) mixture of two similar but > distinct oligonucleotides. The resolution is medium to low (2.7-2.8) but > for a few of these there are some hints from the density for two different > bases at the same position. I'm curious what the best way to approach > refinement would be in this case. Alternate conformation doesn't really > work since the residues have different nucleobases. Having two complete > chains with 0.5 occupancy is overkill since there are only 2 (or 4) > positions in which the sequence differs. I tried just adding a second chain > for the varying residues at 0.5 occupancy and adding link records to the > original chain, however this doesn't seem to respect geometry of the > phosphodiester for the flanking residues. I would appreciate suggestions or > any examples in the PDB that might set me in the right direction. > > --paul >
