Hi,
> I am trying to understand why these classifications happens - to try to
> understand what happened, i would like to get from RELION the projections
> parameters and the score (probability) for each particle against both
> classes during the 3D classifications.
> I am somewhat familiar with the RELION source code but I could not find an
> easy way to print these parameters.
> If someone could point me toward the right area of the codes / or knows
> how to get these parameters, I would greatly appreciate it.
This is not trivial to do and you have to hack ml_optimier.cpp.
Class probability of each particle:
A relevant variable for non-accelerated CPU code is
thr_wsum_pdf_class in MlOptimiser::storeWeightedSums().
Projection parameters for each particle for each class:
We don't have this, because poses are marginalised.
If you want the most likely poses for each class, you have to
take maximum for each class yourself. Again, good place to implement
this is storeWeightedSums.
Best regards,
Takanori Nakane
########################################################################
To unsubscribe from the CCPEM list, click the following link:
https://www.jiscmail.ac.uk/cgi-bin/webadmin?SUBED1=CCPEM&A=1
|