Focus Documentation

Merging modus

This allows to define if 2dx_merge should operate in the “2D” or the “3D” modus. In the 2D modus, available data from non-tilted images are merged into a two-dimensional dataset. In the “3D” modus, the merged dataset has to include data from images of tilted specimen. In this case, 2dx_merge also runs the lattice line fitting procedure, and the final result is a 3D volume.

Lattice Determination

This defines the dimensions of the real-space unit cell. These values should be the same for the entire processing project, and are synchronized between the 2dx_image calls.

Common Image Processing

The merging of the data is done with data from within these boundaries.

Specifies the resolution upto which the merged dataset can/should be used as reference for unbending of the images. The other parameter resmax “Upper Resolution Limit” is used to define the resolution limit up to which the processing of one image, or up to which the merging of all available data should be done. This parameter here “MergeResolution - Resolution of the merged dataset” instead defines, to which resolution the final merged dataset should be limited, to be used as reference. You could for example process data up to 7A resolution, but provide a reference that only contains data up to 10A resolution, so that more noisy data in the 10-7A range are not used to wrongly influence the re-processing of your images.

This defines the vertical Z-direction dimension of the unit cell volume to be reconstructed. In Angstroems. Usually, for a single-layer membrane protein crystal, this could be 200.0 Angstroems. For a double-layered crystal, this can be set to higher values. Have a look under Resolution

Specify the number of lines in the final resultion table of 2dx_avrgamphs (a cosmetic feature for output only)

Specify the highest resolution in the final resultion table of 2dx_avrgamphs (a cosmetic feature for output only). Set to 0 for automatic determination.

Merging

Switch, deciding if a new merged dataset (in 2D or 3D) should be re-created, overwriting the old version. If this is set to no, then the old one is kept, if it exists. You can use this to protect a good reference, once you hove it.

These two parameters define the range and stepsize of the search for the phase origin of one image with respect to the currently available merged dataset. The Origtilt program can be used to merge all available data into one 2D or 3D merged dataset. This is then used as reference, onto which the available image data are aligned. The correct alignment is described as “Phase Origin” for an image, whereby for example a shift of half a unit cell would be described as a shift by 180 deg. If you define to run 61 steps of 6.0 degree step size, then you would scan the entire unit cell range, from -180.0 to +180.0 degrees shift. Once you have the impression that all images are more or less aligned, you can narrow the search range, by for example defining 61 steps of 0.2 degrees step size. That would then perform a search over the area of -3.0 to +3.0 degrees.

Defines which data should be used as reference for the cyclic refinement of the phase origin. The merge.aph file is more robust for early stages of the project. The interpolated lattice lines might be better for well-refined structures.

There are 10 registers for merging results. Register 0 is always the result of the last merging process. Registers 1-9 are storage registers. Choose, which one should be used as reference for refinement of parameters.

There are 10 registers for merging results. Register 0 is always the result of the last merging process. Registers 1-9 are storage registers. Choose, where you want register 0 to be copied to.

Switch, deciding if the merged dataset (works only in 3D) should be used to create a projection map in the same tilt direction as the corresponding image is. This reference image can then be compared with the experimental image (final_map.mrc), to see if the two correspond to each other. This needs the MTZ file format for the merged reference.

This switch allows you to define, if potentially existing 2dx_image results from the Maximum Likelihood (ML) processing of individual images should be used for the merging step, instead of the 2dx_image results from Fourier filtering. Even with this switch to “ML”, the ML data are only used if the individual images allow it. You can decide this for the individual images with the switch “Use Single Particle Result for merging in 2dx_merge” in 2dx_image.

Set this switch to yes to generate PostScript files. This will slow down the processing though.

This defines, which data should be included in the merging process. With a setting of “5”, all spots with IQ values of 6 or worse would not be included in the merging, which can reduce the noise content in the merging process if you have few images to be merged. If you have a very large number of images, you can set this parameter to higher (e.g. “6”) values.

Specifies the highest miller index of spots to be included in the phase origin refinement process, which is done by the MRC program origtilt. A value of 20 is ok in most cases.

Defines if every reflection should be written into the log files during origtiltk runs, i.e. during merging and refinement runs.

Defines the relative weights for phases with respect to amplitudes, for lattice line generation. If your Defocus is determined very precisely i.e. the Thon rings are well placed and therefore the amplitudes are more reliable. In that case, you can use a smaller value (e.g. 2.0). If the defocus values are not too reliable, then a larger value of 3.0 or 4.0 might work better.

IWF, IWP are flags to use individual weights for F's and Phases in input file: -1 = use individual sigmas (1/sigma**2); 0 = set all weights = 1.0; 1 = use individual weights

3D Merging (these do not apply for 2D merging)

This defines the vertical Z-direction dimension up to which the merging or 3D volume reconstruction should be performed. The highest possible z-star value is 0.5, which is the Nyquist frequency in the vertical direction. If you want to apply a resolution limitation in the vertical direction, then you can set this value to something smaller, e.g. 0.2 or 0.05. See also Zstar

Defines, if Lattice Lines should be fitted using F&Phase, or only based on the Intensity data.

Defines the vertical z* interval for plotting (in reciprocal Angstroems). If set to zero, then nothing is plotted.

Defines the real space sampling interval for lattice line profile function, in Angstroems.

Defines the vertical real space envelope that will surround the reconstruction. This should be the thickness of your crystal layer, in Angstroems.

Defines the Half-Width at Half Maximum of the Gaussian falloff of the vertical real-space envelope, in Angstroems.

Defines if a linear or Gaussian falloff of the vertical real-space envelope should be applied. Negative values concern a linear falloff, positive values a Gaussian. In Angstroems.

Defines how coarse or fine the data should be sampled to come up with a first lattice line guess. This is in reciprocal Angstroems, meant along the vertical lattice lines.

Number of refinement cycles to perform. If negative, then output initial guess. If set to zero, then minimizer goes till end. If positive, then it defines the number of cycles to go (25-50 is quite reasonable).

If positive, then print the var/covar matrix. If set to zero, then don't print it.

Merging Refinement

The factor for including the image into the global merge. 1.0 would be normal.

The phase residual during the merging process. The smaller the phase residual, the better the merged image.

Phase error should be smaller than 90.0 - 90.0/SQRT(N), with N being the number of measurements that are averaged.