Train junk detector trained on detergent micelles?

Hi,

I love the micrograph junk detector! I know have a dataset where the protein was frozen with CHAPSO (not a membrane protein), and see detergent micelles in some micrographs. I wondered if it is possible to train the junk detector to learn to identify these? They have similiar dimensions as my protein so might mess with picking…

Thanks in advance,

Matthias

I think this is a better job for heterogeneous refine. Make an ab initio map of some particles that are detergent. Feed this as a decoy along with your protein map into het refine. It will get rid of them.

5 Likes