Picking a box size & pixel size for particle processing - cianfrocco-lab/Old-school-processing GitHub Wiki

[Home]] > [Best practices > Picking box size & pixel size

##Picking a box size & pixel size for particle processing

When it comes to analyzing single particle EM data, there are a few considerations required UP FRONT regarding the pixel size of the data that you will be collecting & the box size of the extracted particles.

###Pixel size Pixel size refers to the size of each pixel within the recorded image of your sample. Typical values will range from 1 Angstrom/pixel to 5 Angstroms/pixel.

####Tradeoff: pixel size vs. imaging area The smaller your pixel size, the more high resolution information will be recorded in your image. However, you will subsequently be imaging a smaller area, and thus get less particles per image.

Therefore, you need to consider the following questions:

• What resolution do you expect that you will achieve with this dataset?
  • Generally, the resolution limit of your dataset will be 2-3X the pixel size (best case!).
  • So, if you are expecting a 15 Angstrom negative stain structure, a pixel size of 2 - 4 Angstroms/pixel is OK because the limit will be 8 - 12 Angstroms.
• Is this dataset being collected using CCD or Direct Detector?
  • The information transfer within a CCD is much less than that of the latest generation of direct electron detection devices. This means that resolution limits for CCDS are worse, being limited to 4 - 5X the pixel size.
  • If you are using a direct detector, then the resolution limit will be 2X the pixel size.

###Box size

When selecting a box size that will be used for your particle extraction, a good rule of thumb is:

Box size = 2.5 x Particle diameter

This gives the particles to move within the box during refinement and analysis.

If you'd like a rigorous definition of box size, make sure to check out page 726 in Rosenthal & Henderson (2003) JMB.