Following on the previous thread on Fe-SAD, there is a solution(?) from
autoSHARP with an OK-ish map and model. A quick refinement with phenix.refine
gave me an R/Rfree of 28/31 % (2 A, although for the phasing run I told
autoSHARP to use only up to 2.5 A). However, the model was filled in with about
6000 dummy atoms (DUM) plus some regions with just Gly. I did provide a
sequence file, so not sure why it didn't build them in. When I deleted all the
DUM atoms, the R/Rfree increased to about 48 %.
I am wondering, are the positions of DUM atoms correct and I just need to build
it manually somehow?