Hi Eike,
It could be overfitting. You will have to assess yourself whether the
quality of the images keeps improving. Most data sets converge quite OK
in 25 iterations.
HTH,
Sjors
On 03/28/2017 02:50 PM, Schulz, Eike-Christian wrote:
> Dear all,
>
> I am working on low resolution data with Relion2.0. Since I am new to this field I am not sure for how long it is meaningful to run the 2D classification ?
>
> Since my particle is highly symmetric I initially run 25 cycles with an angular sampling of 5 degrees and then another 25 cycles at 2.5 degrees. However, when I plot the Current-Resolution output over the run-numbers I still see a slight improvement in resolution.
>
> Is this still useful or already overfitting ?
> How do you decide when to stop ?
>
> Best regards
>
> Eike
>
> ++++ Executing new job on Mon Mar 27 16:31:27 2017 ++++ with the following command(s):
> `which relion_refine_mpi` --o Class2D/job018/run --i Extract/job016/particles.star --preread_images --pool 3 --iter 25 --tau2_fudge 2 --particle_diameter 1000 -
> -K 150 --flatten_solvent --zero_mask --strict_highres_exp 30 --oversampling 1 --psi_step 12 --offset_range 5 --offset_step 2 --norm --scale --j 10 --gpu "0:1"
>
> ++++ Executing new job on Mon Mar 27 17:55:43 2017
> ++++ with the following command(s):
> `which relion_refine_mpi` --continue Class2D/job018/run_it025_optimiser.star --o Class2D/job018/run_ct25 --preread_images --pool 3 --iter 50 --tau2_fudge 2 --pa
> rticle_diameter 1000 --oversampling 1 --psi_step 5 --offset_range 5 --offset_step 2 --j 10 --gpu "0:1"
> ++++
>
--
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
|