[phenixbb] polder vs fem vs composite omit map vs maximum entropy

Pavel Afonine pafonine at lbl.gov
Thu Jan 19 20:27:25 PST 2017


> So I got a MR solution and maybe a dozen residues are not clearly 
> defined (most of them in the surface)

Sounds like you are in the beginning of structure solution/refinement.

> I used to compute the feature enhanced map but now I can see there are 
> a lot of options in phenix->maps. Which one is better for me? (~130 
> residues, 1.5 A resolution, ~ 90% complete)
> Probably I should try them all, and compare them visually in coot with 
> appropiate sigmas (taken from maps-> map comparison) (is that a good 
> workflow? what you guys do?) ...but some of the maps take a lot of 
> time to compute. So any help will be appreciated.

"(Rebuild->Refine) * N_times" until you are done fixing gross errors; 
maps will be supplied to you be refinement engine each time cycle is 
over; no need to use "phenix->maps".

Then towards the end focus on smaller details. This is when FEM and 
Polder maps will be most helpful.

Make sure you fully understand what each map calculation option does and 
when it is best to use it. Ask specific questions if not sure. The 
strategy "try them all and see what happens" is certainly good in some 
cases (like when trying plausible refinement choices) by does not sound 
particularity great in your case.

Pavel



More information about the phenixbb mailing list