[phenixbb] scattering_table in phenix.real_space_refine

[Pavel Afonine]

Hi Keitaro,

this sounds like a bug, thanks for reporting!
Could you please send me the code diffs off list?

Thanks,
Pavel

On 3/25/19 07:04, Keitaro Yamashita wrote:
> Dear PHENIX developers,
>
> For the refinement of CryoEM structure, I changed scattering_table to
> electron from default (n_gaussian in phenix-1.13), but the resultant
> coordinates and ADPs were exactly same while map-model CCs were
> improved. Looking into the code, I noticed scattering_table is not
> passed to the ADP refinement code
> (mmtbx.refinement.real_space.adp.real_space_group_adp_refinery_via_reciprocal_space)
> and xray_structure is newly constructed from a pdb_hierarchy object;
> therefore a default table in
> scattering_type_registry() (which is n_gaussian) is always used. Is
> this an intended behavior?
>
> I modified the code to use the scattering table of electron in ADP
> refinement, and CCs were improved (CC_mask 0.7566 to 0.7813; both
> evaluated with scattering_table=electron).
>
> Best regards,
> Keitaro
> _______________________________________________
> phenixbb mailing list
> phenixbb at phenix-online.org
> http://phenix-online.org/mailman/listinfo/phenixbb
> Unsubscribe: phenixbb-leave at phenix-online.org