Table column convention

From Dynamo
Jump to navigation Jump to search


Here we list the significance of each column in a table. Each row represents a single particle.

col 1: tag

Identifies the particle number, so that each particle in the data folder indexed by the table is identified.

col 2: align

0 or 1 Used when a table is fed into Dynamoas initial table for an alignment project. Value 1 tells Dynamo that the particle is to be used during an alignment project, values 0 skips the particle. hola

col 3: average

0 or 1 Written by Dynamo during an alignment project, indicating if the particle was used or not for the final average. A particle can be excluded from the average for different reasons, mainly because its cc score didn't qualify it to pass the threshold selected by the user.


col 4:6: shifts

In pixels, from the center of the box. In a table fed into an alignment project, the user can use these columns if an approximate a priori knowledge of the shifts is available. In the refined table output by Dynamo, cols 4:6 simply show the results of the alignment.

col 7:9: euler angles

In a table fed into an alignment project, the user can use these columns if an approximate a priori knowledge of the angles is available. This is the case when the table has been produced with some geometric model that captures initial orientations (like helices, filaments or membranes). In such cases. In the refined table output by Dynamo, cols 7:9 simply show the results of the alignment.

col 13: ftype (type of fourier sampling)

Describes the fraction of the Fourier space covered by the particle. Basically it represents the imaging geometry.

  • 0 or 'full'
  • 1 or 'single' beam along z, tilt around y
  • 2 or 'singlex' beam along z, tilt around x
  • 3 or 'cone', beam along z
  • 4 or 'double', 'dual' beam along z, tilts around y and x
  • 5 or 'custom'
  • 6 or 'tilt_z_beam_y' beam along y, tilt around z
  • 7 or 'tilt_x_beam_y' beam along y, tilt around x

For total characterization of the Fourier components covered by the particle, the information in this column needs to be complemented with other columns [14 to 19] depending on the ftype.

ftype 1: conventional tilt about y

It requires the user to fill the columns 14 and 15 with the minimum and maximal tilt angles around y. Other columns intended for fourier wedge description (columns 16 to 19) are simply ignored.

ftype 2: tilt about x

It requires the user to fill the columns 16 and 17 with the minimum and maximal tilt angles around x. Other columns intended for fourier wedge description (columns 114,15, 18 and 19) are simply ignored in this case.

ftype 5: custom

When a particle is indicated to have a custom mask, then the data folder needs to include a volume file with zeros and ones that describes this custom geometry. This is useful for instance when the user wants to explicite the different Fourier planes sampled during tilt colection, instead of describing them like a wedge. The file representing this fsampling should be called pfmask, i.e. pfmask_00010.em for the corresponding particle particle_00010.em and so on.

ftype 6 and 7

It is not advisable to use these geometries when creating tomograms. Slices orthogonal to the direction of the beam are more difficult to interpret visually. This makes useful to keep the direction z as the direction of the beam, because a slice in z is kept together in the disk and also in memory (check this) and is more readily accessible when exploring a tomogram.

cols 14:15: y-range of missing wedge

cols 24-26: position in tomogram

Given in pixels.

col 34: reference

This column is written by Dynamo during the computation of a multireference alignment project. It marks the reference channel that yielded the highest score for the particle. A "reference" can be seen as a "class" produced by multireference alignment and classification.

col 35: subreference

This column is typically written by Dynamo while creating averages of the particles assigned to different clusters computed through PCA+Kmeans or Hierarchical Ascending Clustering. Thus, in Dynamo jargon, "subreference" can be seen as a "class" produced by classification through PCA+ Kmeans or HAC. The name "subreference" comes from the fact that we frequently use PCA analysis on top of the results attained by Multireference Analysis.

col 41 onwards: eigencomponents

Produced by Dynamowhen creating an eigentable during PCA