Hi Leo,
There currently does not exist an alternative for what you propose, and
it's not coming out any time soon (we're already trying to get relion-2.0
stable in-house). You could try and modify src/ml_optimiser.cpp to make
the change you propose.
Best,
Sjors
> Hi Sjors and others, I wonder if there is such a possibility or if it is
> possible to add it in Relion.
>
> Sometimes from 3D classification run the results look apparently better
> (estimated accuracy is better) than from 3D autorefine. However, there
> seems to be no way to get proper unfiltered map out of it. With
> "relion_reconstruct" on _data.star only single particle orientation is
> used and resolution suffers, as I also understand from previous
> discussions.
>
> So if there was an option to have "continue old run" for 'converged' 3D
> classification with an option to set output maps to unfiltered, that would
> seem to solve the problem and should be easy to implement. Or is there may
> be some other workaround?
>
> Thanks for any help!
>
>
>
--
Sjors Scheres
MRC Laboratory of Molecular Biology
Francis Crick Avenue, Cambridge Biomedical Campus
Cambridge CB2 0QH, U.K.
tel: +44 (0)1223 267061
http://www2.mrc-lmb.cam.ac.uk/groups/scheres
|