[phenixbb] phenix.resolve_cryo_em resolution and real space refinement
David Lawson (JIC)
david.lawson at jic.ac.uk
Wed Oct 14 05:05:44 PDT 2020
Thanks very much for the clarification!
From: Tom Terwilliger <tterwilliger at newmexicoconsortium.org>
Sent: 14 October 2020 12:52
To: David Lawson (JIC) <david.lawson at jic.ac.uk>
Cc: phenixbb at phenix-online.org; Tom Terwilliger <tterwilliger at newmexicoconsortium.org>
Subject: Re: [phenixbb] phenix.resolve_cryo_em resolution and real space refinement
Yes, the 7.08 in:
FSCref=0.5: 7.08 6.69 6.63
Is referring to the estimated initial resolution where the half-dataset correlation is 0.143, leading to an estimated correlation to the true map of 0.5. The 6.69 number is the map-phasing map, and the 6.63 is the density modified (combined) map.
Yes, the density modification resolution limit is chosen empirically as a finer resolution than you supply as a nominal resolution.
I think that your real-space refinement correlations won't be very different if you use 7.1 or 6.6 A as your nominal resolution. I'd suggest using the estimated resolution from density modification but it may be more important to be consistent with other runs if you want to compare results.
All the best,
On Wed, Oct 14, 2020 at 5:29 AM David Lawson (JIC) <david.lawson at jic.ac.uk<mailto:david.lawson at jic.ac.uk>> wrote:
I have run phenix.resolve_cryo_em using half maps at 7.5 A resolution (the postprocessed map is at 7.1 A) and I just wanted to clarify a couple of things (I have pasted the end of the logfile below this message). I set the resolution to 7.52 A but didn’t supply a resolution for density modification – it seems to have chosen a value of 5.671 A
(1) It states “Estimated improvement of -0.450 A to d_min(FSCref=0.5): 6.628 A” suggesting to me that it agrees with the postprocessed resolution. Is this the value of 7.08 on this line: “FSCref=0.5: 7.08 6.69 6.63”?
(2) I understand that the actual process of real space refinement doesn’t care about the resolution of the map, but you need to supply it to get meaningful CC values. So my question is what resolution should I supply to RSR if I use the density modified map?
I realise that a value of “-0.450 A” is at the upper end of what to expect from density improvement, but I can report that the DM map does look noticeably better than the postprocessed map, especially in regions that are less well defined.
Many thanks in advance for any advice
FSCref=0.5: 7.08 6.69 6.63
Estimated improvement in d_cc_star=0.5: -0.450 A
This is new current best working_map_coeffs_info. Estimated
improvement of -0.450 A to d_min(FSCref=0.5): 6.628 A
Done testing....R-value: 0.530 Try: 1 ['database 5', 'resolution: 5.67 ', 'dm_resolution: 5.67 ', 'density_modify_unsharpened_maps:False', 'apply_cc_star:True', 'b_iso: None A**2']
Best density modification cycle 1 is try 1 with
Methods used: ['database 5', 'resolution: 5.67 ', 'dm_resolution: 5.67 ', 'density_modify_unsharpened_maps:False', 'apply_cc_star:True', 'b_iso: None A**2']
Estimated improvement of -0.450 A to d_min(FSCref=0.5): 6.628 A
Result: -0.450 A to d_min(FSCref=0.5): 6.628 A R-value: 0.530
Wrote working map to /home/lawsond/phenix/ResolveCryoEM_1/map_denmod.ccp4
SUMMARY AFTER CYCLE OF DENSITY-MODIFICATION
Overall best density modification Cycle 1 try 1
Estimated improvement: -0.450 A to d_min(FSCref=0.5): 6.628 A R-value: 0.530
Denmod resolution: 5.671 A
Database: 5.000 A
Writing out final map files…
Prof. David M. Lawson
Department of Biological Chemistry,
John Innes Centre,
NR4 7UH, UK.
Email: david.lawson at jic.ac.uk<mailto:david.lawson at jic.ac.uk>
phenixbb mailing list
phenixbb at phenix-online.org<mailto:phenixbb at phenix-online.org>
Unsubscribe: phenixbb-leave at phenix-online.org<mailto:phenixbb-leave at phenix-online.org>
Thomas C Terwilliger
Laboratory Fellow, Los Alamos National Laboratory
Senior Scientist, New Mexico Consortium
100 Entrada Dr, Los Alamos, NM 87544
Email: tterwilliger at newmexicoconsortium.org<mailto:tterwilliger at newmexicoconsortium.org>
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the phenixbb