Bumpy FSC curves

Vruiz · September 27, 2024, 3:45pm

Hi, y’all!

I’ve noticed weird FSC curves using either Reconstruct only and Non-uniform refinement jobs after importing a subset of refined particles from Relion to cryosparc, and I’d like to know what’s your opinion:

The same subset of particles gives a normal FSC in relion with a modest resolution:

I imported these particles as:
Particle meta path: Relion/3DRefine/run_data.star
Particle data path: Relion/Extract/extracted particles folder from relion

The particles were extracted using float-16 in Relion.

Using the job Reconstruct only or Non-uniform refinement yields a medium-resolution model. 2D classification yields good 2D averages. I get the same FSC if I use the output particles from the 2D for Non-uniform refinement.

Here is the particle import job output:

License is valid.

Launching job on lane default target c105627.dhcp.swmed.org ...

Running job on master node hostname c105627.dhcp.swmed.org

[CPU:   78.6 MB  Avail: 377.29 GB]
Job J163 Started

[CPU:   78.7 MB  Avail: 377.29 GB]
Master running v4.6.0, worker running v4.6.0

[CPU:   78.7 MB  Avail: 377.29 GB]
Working in directory: /data1/data/J163

[CPU:   78.7 MB  Avail: 377.29 GB]
Running on lane default

[CPU:   78.7 MB  Avail: 377.29 GB]
Resources allocated: 

[CPU:   78.7 MB  Avail: 377.29 GB]
  Worker:  c105627.dhcp.swmed.org

[CPU:   78.7 MB  Avail: 377.29 GB]
  CPU   :  [0]

[CPU:   78.7 MB  Avail: 377.29 GB]
  GPU   :  []

[CPU:   78.7 MB  Avail: 377.29 GB]
  RAM   :  [0, 1, 2]

[CPU:   78.7 MB  Avail: 377.29 GB]
  SSD   :  False

[CPU:   78.7 MB  Avail: 377.29 GB]
--------------------------------------------------------------

[CPU:   78.7 MB  Avail: 377.29 GB]
Importing job module for job type import_particles...

[CPU:  264.0 MB  Avail: 377.15 GB]
Job ready to run

[CPU:  264.0 MB  Avail: 377.15 GB]
***************************************************************

[CPU:  264.0 MB  Avail: 377.15 GB]
Importing particles from  /media/ExternalBackup6/Relion/Refine3D/job209/run_data.star

[CPU:  264.3 MB  Avail: 377.15 GB]
File extension is  star

[CPU:  264.3 MB  Avail: 377.15 GB]
Importing star file.

[CPU:  277.3 MB  Avail: 377.14 GB]
--------------------------------------------------------------

[CPU:  277.3 MB  Avail: 377.14 GB]
Loaded star file with 7888 items

[CPU:  277.3 MB  Avail: 377.14 GB]
Fields loaded from star file:  ['rlnOpticsGroupName', 'rlnOpticsGroup', 'rlnMicrographOriginalPixelSize', 'rlnVoltage', 'rlnSphericalAberration', 'rlnAmplitudeContrast', 'rlnImagePixelSize', 'rlnImageSize', 'rlnImageDimensionality', 'rlnCtfDataAreCtfPremultiplied', 'rlnCoordinateX', 'rlnCoordinateY', 'rlnAutopickFigureOfMerit', 'rlnClassNumber', 'rlnAnglePsi', 'rlnImageName', 'rlnMicrographName', 'rlnCtfMaxResolution', 'rlnCtfFigureOfMerit', 'rlnDefocusU', 'rlnDefocusV', 'rlnDefocusAngle', 'rlnCtfBfactor', 'rlnCtfScalefactor', 'rlnPhaseShift', 'rlnGroupNumber', 'rlnAngleRot', 'rlnAngleTilt', 'rlnOriginXAngst', 'rlnOriginYAngst', 'rlnNormCorrection', 'rlnLogLikeliContribution', 'rlnMaxValueProbDistribution', 'rlnNrOfSignificantSamples', 'rlnRandomSubset', 'rlnImageOriginalName']

[CPU:  277.3 MB  Avail: 377.14 GB]
--------------------------------------------------------------

[CPU:  277.3 MB  Avail: 377.14 GB]
Reading particle data locations...

[CPU:  277.3 MB  Avail: 377.14 GB]
  Reading rlnImageName to get indices and paths..

[CPU:  279.9 MB  Avail: 377.14 GB]
  Warning: Parameter particle_blob_path was set, which overrides rlnImageName despite the latter being present in the input star file.

[CPU:  279.9 MB  Avail: 377.14 GB]
  Parameter particle_blob_path was set and is a directory, so will be used as the search base for finding referenced data paths in the star file.

[CPU:  279.9 MB  Avail: 377.14 GB]
Searching for linked data files...

[CPU:  279.9 MB  Avail: 377.14 GB]
rlnCtfScalefactor is present but all values are unity; these will be ignored.

[CPU:  279.9 MB  Avail: 377.14 GB]
--------------------------------------------------------------

[CPU:  279.9 MB  Avail: 377.14 GB]
Compiling CTF information...

[CPU:  279.9 MB  Avail: 377.14 GB]
Warning: No exposures were connected as inputs, so correspondences cannot be found, so output will not contain pick locations for particles.

[CPU:  279.9 MB  Avail: 377.14 GB]
--------------------------------------------------------------

[CPU:  279.9 MB  Avail: 377.14 GB]
Compiling particle pose information...

[CPU:  279.9 MB  Avail: 377.14 GB]
  Converting euler angles..

[CPU:  279.9 MB  Avail: 377.14 GB]
  Converting rlnOriginXAngst to pixels using parameter psize_A

[CPU:  280.1 MB  Avail: 377.14 GB]
--------------------------------------------------------------

[CPU:  280.1 MB  Avail: 377.14 GB]
Compiling particle pick statistics...

[CPU:  280.1 MB  Avail: 377.14 GB]
--------------------------------------------------------------

[CPU:  280.1 MB  Avail: 377.14 GB]
Particle information has now been imported for 7888 particles, creating outputs...

[CPU:  280.4 MB  Avail: 377.14 GB]
Found references to 4003 unique data files

[CPU:  281.7 MB  Avail: 377.14 GB]
Import paths were unique at level -1

[CPU:  281.7 MB  Avail: 377.14 GB]
Example imported relative path:
 J163/imported/005020243756454551131_PN1_113-11_00001_X+0Y+0-1.mrcs

[CPU:  281.7 MB  Avail: 377.14 GB]
Reading MRC file header to check shape...

[CPU:  283.4 MB  Avail: 377.14 GB]
  Setting validator for blob to: True

[CPU:  283.4 MB  Avail: 377.14 GB]
  Setting validator for ctf to: True

[CPU:  283.4 MB  Avail: 377.14 GB]
  Setting validator for location to: False

[CPU:  283.4 MB  Avail: 377.14 GB]
  Setting validator for alignments3D to: True

[CPU:  283.4 MB  Avail: 377.14 GB]
  Setting validator for filament to: False

[CPU:  283.4 MB  Avail: 377.14 GB]
  Setting validator for pick_stats to: True

[CPU:  283.4 MB  Avail: 377.13 GB]
--------------------------------------------------------------

[CPU:  283.4 MB  Avail: 377.13 GB]
Loaded 7888 particles.

[CPU:  283.4 MB  Avail: 377.13 GB]
  Common fields: 

[CPU:  283.4 MB  Avail: 377.13 GB]
          ctf/accel_kv :  {300.0}

[CPU:  283.4 MB  Avail: 377.13 GB]
             ctf/cs_mm :  {207.0}

[CPU:  283.4 MB  Avail: 377.13 GB]
      ctf/amp_contrast :  {0.1}

[CPU:  283.4 MB  Avail: 377.13 GB]
      ctf/exp_group_id :  {1}

[CPU:  283.4 MB  Avail: 377.13 GB]
          blob/psize_A :  {1.65}

[CPU:  283.4 MB  Avail: 377.13 GB]
            blob/shape :  [256 256]

[CPU:  283.4 MB  Avail: 377.13 GB]
--------------------------------------------------------------

[CPU:  283.4 MB  Avail: 377.13 GB]
Making plots...

Lowpass Filtered Images:

[png]
[pdf]
J163_lowpass_filtered_images.png
Raw Images:

[png]
[pdf]
J163_raw_images.png
[CPU:  370.9 MB  Avail: 377.04 GB]
--------------------------------------------------------------

[CPU:  370.9 MB  Avail: 377.04 GB]
Compiling job outputs...

[CPU:  370.0 MB  Avail: 377.04 GB]
Updating job size...

[CPU:  370.0 MB  Avail: 377.03 GB]
Exporting job and creating csg files...

[CPU:  370.0 MB  Avail: 377.03 GB]
***************************************************************

[CPU:  370.0 MB  Avail: 377.03 GB]
Job complete. Total time 8.21s

I appreciate any feedback

Best
Victor

olibclarke · September 27, 2024, 3:48pm

Either (1) incorrect input Cs or pixel size or (2) very severe beam tilt. Former more likely, so check you didn’t make an error when importing the data, for the latter you can try global CTF refinement to estimate/correct beam tilt.

Vruiz · September 27, 2024, 4:11pm

Hi olibclarke

I found an error in the CS import particles job: The spherical aberration was set to 270 instead of 2.70. I removed all microscope parameters and let CS read them from the .star file, this solved the problem.

Thank you!