If I am not sure of the gain reference (flipped or not, and if so, on Y or X-axis). How can I know by trial and error? *If it was not very distinguished as it was in the video tutorials… ?
When using Falcon EER, the gain reference is in the same native orientation. The same goes for Gatan cameras, raw data and gain reference always are in native orientation, no rotation or flip. Things might be different if some file conversion has been done on the raw data. SerialEM uses a rotation/flip tag in the TIF or MRC header (R/F).
If you are using Gatan K3, then the operator should give you the information of gain flipping/rotation.
From the micrograph analysis, It comes with experience. After doing motion correction your micrographs should be clean of any artefacts, like horizontal/vertical lines. This says the rotation/flipping is correct.
Hi Nihal,
Not flipping the gain correctly can result in artifacts that appear as lines on micrographs. You can generally see it right off the bat during import, but if you miss it you’ll catch it during curation. Here are a few examples where the gain was not flipped in Y. Notice the line that appears to divide the micrograph in two through the middle. If you see the following on your micrographs or something similar you’ll need to flip the gain.
Relion and MotionCor/Aretomo flip TIFFs after they are read. So usually you have to flip the gain when use those programs, and not when you use cryoSPARC. (For Gatan cameras, this includes both .mrc gain copied by SerialEM or the .dm4 gain converted by dm2mrc or e2proc2d.py).
However, if you use relion_estimate_gain on your TIFFs then the result does need to be flipped in cryoSPARC (and then not in Relion).
I am not aware of any circumstance when rotating the gain or flipping in X would be necessary, unless one has separately transformed the images somehow.