[3dem] Model Refinement

[Mohamed A Sayed]

Hello
I am new to cryo-em map reconstruction and want to ask for advice about improving the resolution of the final model. Please see below the steps that I applied.
I think my model current resolution is limited by my ability to perform proper refinement.

- Is there a way to improve the refinement and to check if this best possible final resolution obtained by the experimental data?

- At such resolution, is it worth performing molecular replacement since there is a pdb file already available from x-ray?
Thank you very much for your help.

I am using Relion 3.0 to performing Bayesian polishing, 3d auto-refine and generating mask on 83,000 particles, the final resolution came to be 4.27 A
mask paramters:
lowpass filter map 15 A
Initial binarisation threshold: 0.003
extend binary map this many pixels:0
add a soft-edge of this many pixels: 6

== Masking input maps ...
  + randomize phases beyond:       8.28125 Angstroms
== Dividing map by the MTF of the detector ...
  + mtf STAR-file:                 mtf2_k2_300kV.star
== Applying sqrt(2*FSC/(FSC+1)) weighting (as in Rosenthal & Henderson, 2003) ...
== Fitting straight line through Guinier plot to find B-factor ...
  + fit from resolution:           10
  + fit until resolution:          0
  + slope of fit:                  -43.0451
  + intercept of fit:              -12.7354
  + correlation of fit:            0.908274
  + apply b-factor of:             -172.18
== Low-pass filtering final map ...
  + filter frequency:              4.27419
== Writing output files ...
  + Processed map:                 PostProcess/job064/postprocess.mrc
  + Processed masked map:          PostProcess/job064/postprocess_masked.mrc
  + Metadata file:                 PostProcess/job064/postprocess.star
  + FINAL RESOLUTION:              4.27419


I made another 3D classification on the shiny particles and selected one 3D class of 44,000 particles of the highest resolution performed 3d auto-refine, applied same mask and did post-processing my final resolution became 4.42 A

 + input mask:                    MaskCreate/job057/mask.mrc
== Masking input maps ...
  + randomize phases beyond:       8.0303 Angstroms
== Dividing map by the MTF of the detector ...
  + mtf STAR-file:                 mtf2_k2_300kV.star
== Applying sqrt(2*FSC/(FSC+1)) weighting (as in Rosenthal & Henderson, 2003) ...
== Fitting straight line through Guinier plot to find B-factor ...
  + fit from resolution:           10
  + fit until resolution:          0
  + slope of fit:                  -49.6022
  + intercept of fit:              -12.9622
  + correlation of fit:            0.970846
  + apply b-factor of:             -198.409
== Low-pass filtering final map ...
  + filter frequency:              4.41667
== Writing output files ...
  + Processed map:                 PostProcess/job078/postprocess.mrc
  + Processed masked map:          PostProcess/job078/postprocess_masked.mrc
  + Metadata file:                 PostProcess/job078/postprocess.star
  + FINAL RESOLUTION:              4.41667

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.ncmir.ucsd.edu/pipermail/3dem/attachments/20181205/e627786e/attachment.html>