Hi folks,
I have been having difficulties getting nice 2D class averages in RELION.
When I set "NO" to the option "Ignore CTFs until first peak" my 2D classes look low resolution/featureless. If the option is set to "YES" then the 2D classes look high resolution/with internal features, but have dark rings around them that look like the point spread function and have rays going outward from the center of the box, which most likely related to overfitting/noise bias. I tried to use a higher T value (e.g 3) when the "Ignore CTFs until first peak" is set to "NO" or include the Fourier components only up to 15A resolution with the "Ignore CTFs until first peak" set to "YES", but nothing helps.
Has anyone of you had a similar issue? What was you solution?
Many thanks!
|