search

leginon-org / leginon-redmine-archive

1 stars 0 forks source link

Write out rlnBeamTiltGroupName column in star file for RELION 3 beam tilt refinement? [Feature Request] #6035

Open leginonbot opened 6 months ago

leginonbot commented 6 months ago

Author Name: Oliver Clarke (@olibclarke) Original Redmine Issue: 6035, https://emg.nysbc.org/redmine/issues/6035 Original Date: 2018-09-20

Hi Anchi,

Relion 3 now supports refining multiple beam tilt groups in a single star file (see: https://bitbucket.org/scheres/relion-3.0_beta/issues/3/beam-tilt-estimation-group-micrographs ).

It does so by use of the rlnBeamTiltGroupName column, which defines the groups of micrographs that should have similar beam tilt values.

Would it be possible to add this column to the summary star file which currently has estimated tilt values? This would allow for very convenient estimation of beam tit for groups of micrographs with nominally similar image shift.

Cheers Oli

leginonbot commented 6 months ago

Original Redmine Comment Author Name: Anchi Cheng (@anchi2c) Original Date: 2018-09-25T06:16:27Z

Oli,

This can be added but it will not give the best result. Within the group there will be beam-image shift variation present because stage movement is not accurate. The better option which you may ask as a Relion feature is to take the beam tilt metadata in the star file and ask relion to shift the center of the beam tilt distribution in the refinement because that is harder to be correct. How much beam tilt is obtained by beam-image shift is fairly stable.

leginonbot commented 6 months ago

Original Redmine Comment Author Name: Oliver Clarke (@olibclarke) Original Date: 2018-09-27T14:45:42Z

Hi Anchi,

That would definitely be better and I will request it - however I would note that splitting the data into image shift groups was used in the Relion 3 paper with good results (see Fig 6 supp 1): https://www.biorxiv.org/content/biorxiv/early/2018/09/19/421123.full.pdf, so while not perfect it may still be helpful I think

Cheers Oli

Anchi Cheng wrote:

Oli,

This can be added but it will not give the best result. Within the group there will be beam-image shift variation present because stage movement is not accurate. The better option which you may ask as a Relion feature is to take the beam tilt metadata in the star file and ask relion to shift the center of the beam tilt distribution in the refinement because that is harder to be correct. How much beam tilt is obtained by beam-image shift is fairly stable.