Question Regarding Files to be Analyzed

Submitted by Unknown on Wed, 2009-12-16 20:39

Forums

ctffind & ctftilt

Hello Dr. Grigorieff!

I'm expecting my pictures to be about 1um underfoucs. And yet the output numbers indicate that they are at most, 100nm underfocus.

Does it make a difference to ctffind that the mrc file it analyzes was originally a tif file?
Since the scanner I'm using cannot save the scanned image in mrc format, I saved it as a tif first; then I used ImageJ to convert that tif file into .mrc format.
Does this process somehow affect the output values? Perhaps information is lost and/or added when the tiff file is converted to a mrc one, as opposed to having the original scan in mrc format?

Thanks!

niko Wed, 2009-12-16 22:11

If the conversion from tiff to mrc was done correctly there should be no loss of data or information. However, sometimes the tiff conversion does not work as expected. Some programs cannot deal with the bit depth of 16bit tiffs, I think.

In any case, you should view the converted mrc file using Ximdisp (MRC image processing package) or another display program to make sure the conversion worked.

When you run ctffind3, also make sure your expected defocus lies within the search range (fourth and fifth parameters on the second input line of ctffind3).

Also make sure you set the low resolution limit (second parameter on the second input line) to a value that will exclude the very low resolution background. You can check that the resolution limits are reasonable by looking at the diagnostic output of the power spectrum and CTF fit. The CTF fit should ideally only include a range where you see Thon rings. The fit can be affected by the strong low resolution ramp/background and by the high resolution noise. Both should therefore be excluded from the fit.

Finally, it always helps if there is some carbon visible in the image. If the image shows only ice the contrast will be quite weak and the Thon rings will be difficult to pick up.

Let me know if this helps.