Basic EM metadata support

[wmoore]

Current proposal:

Most EM metadata/attributes can go at the 'Image' level, in a single 'EM-settings' element, instead of attached to Channel, Plane etc. In this proposal, the EM data model does not attempt to mimic the current LM model, E.g. by attaching LightSource? / to Channel. It would also not add anything under Instrument.

The goal here is simplicity, and follows the approach of current EM data models where there is no attempt to model Instrument, Objectives etc. See

•  http://blake.bcm.edu/emanwiki/Eman2Metadata
•  http://www.wadsworth.org/spider_doc/spider/docs/image_doc.html
•  http://www.imagescience.de/formats/formats.htm

The only limitation is support of formats (E.g. spider) where it is common to bundle many different particle images into a single stack for convenience, making it 'possible' that each plane has different value for a particular attribute (see below). NB. This should also allow for future support of Images with one or more LM channels and one EM channel, even though all the EM data would be at the Image level.

The essential EM image metadata is voltage, magnification, defocus, electron dose, energy filter. Possibly Cs (used by EMAN2). In the EM-DB schema, the units are

• acceleratingVoltage (kv)
• defocus (nm)
• electronDose (e/A**2)
• energyFilter (String)
• coefficient of Spherical Abberation "naminalCs" (mm)

and could be put into a single EM-settings element, on Image.

• Magnification - simply put this in a 'dummy' Objective, OR in an EM-settings element.

Defocus could be at the Image level, since people only take a single defocus per image. But sometimes take 2 images at different defocus values. So maybe this could be another dimension, at the Plane level?! Also, defocus is assigned to particles, and some Spider images are a stack of Particles, each with (potentially) a different defocus? Maybe in practice they are all the same? (see e-mail 13th Jan 2010).

• Tilt angle (degrees)

This should go on plane, and is really another dimension. Needs to wait for N-dimension to be implemented. Increment is not always consistent (Store as PlaneInfo?).

• Pixel sizes might be in Angstroms or Nanometers in EM files. Need to convert these to microns or add support for units: #114.

[wmoore]

Screen-shot of EMAN2 dbd (in e2display.py) showing image attributes.

[wmoore]

Looking at the BDB repository for EMAN2 single-particle workflow using e2display.py. The original images are mrc, and all the MRC attributes are found in the bdb for the original images and all those that are derived from them: E.g. Particles. Particles also have additional attributes such as width, height, source coordinates, max intensity etc. See attached image.

[wmoore]

Text file from SEM image.