Hi Relion users,
I am observing an unusal behaviour while doing 2d classification of
stained data in RELION.
The estimated angular accuracy decreases in the begining and then it
goes to '999'. However, the class averages look more featured (!?!?).
Please see the attachment.
r1_65k_T3_it000_optimiser.star:_rlnOverallAccuracyRotations
999.000000
r1_65k_T3_it001_optimiser.star:_rlnOverallAccuracyRotations
999.000000
r1_65k_T3_it002_optimiser.star:_rlnOverallAccuracyRotations
999.000000
r1_65k_T3_it003_optimiser.star:_rlnOverallAccuracyRotations
999.000000
r1_65k_T3_it004_optimiser.star:_rlnOverallAccuracyRotations
999.000000
r1_65k_T3_it005_optimiser.star:_rlnOverallAccuracyRotations
23.200000
r1_65k_T3_it006_optimiser.star:_rlnOverallAccuracyRotations
11.400000
r1_65k_T3_it007_optimiser.star:_rlnOverallAccuracyRotations
8.000000
r1_65k_T3_it008_optimiser.star:_rlnOverallAccuracyRotations
6.700000
r1_65k_T3_it009_optimiser.star:_rlnOverallAccuracyRotations
5.750000
r1_65k_T3_it010_optimiser.star:_rlnOverallAccuracyRotations
5.450000
r1_65k_T3_it011_optimiser.star:_rlnOverallAccuracyRotations
5.300000
r1_65k_T3_it012_optimiser.star:_rlnOverallAccuracyRotations
4.950000
r1_65k_T3_it013_optimiser.star:_rlnOverallAccuracyRotations
4.750000
r1_65k_T3_it014_optimiser.star:_rlnOverallAccuracyRotations
4.650000
r1_65k_T3_it015_optimiser.star:_rlnOverallAccuracyRotations
4.500000
r1_65k_T3_it016_optimiser.star:_rlnOverallAccuracyRotations
6.800000
r1_65k_T3_it017_optimiser.star:_rlnOverallAccuracyRotations
6.550000
r1_65k_T3_it018_optimiser.star:_rlnOverallAccuracyRotations
8.600000
r1_65k_T3_it019_optimiser.star:_rlnOverallAccuracyRotations
8.600000
r1_65k_T3_it020_optimiser.star:_rlnOverallAccuracyRotations
8.100000
r1_65k_T3_it021_optimiser.star:_rlnOverallAccuracyRotations
7.800000
r1_65k_T3_it022_optimiser.star:_rlnOverallAccuracyRotations
7.500000
r1_65k_T3_it023_optimiser.star:_rlnOverallAccuracyRotations
999.000000
r1_65k_T3_it024_optimiser.star:_rlnOverallAccuracyRotations
999.000000
r1_65k_T3_it025_optimiser.star:_rlnOverallAccuracyRotations
999.000000
r1_65k_T3_it000_model.star:_rlnCurrentResolution
inf
r1_65k_T3_it001_model.star:_rlnCurrentResolution
84.508001
r1_65k_T3_it002_model.star:_rlnCurrentResolution
84.508001
r1_65k_T3_it003_model.star:_rlnCurrentResolution
84.508001
r1_65k_T3_it004_model.star:_rlnCurrentResolution
70.423334
r1_65k_T3_it005_model.star:_rlnCurrentResolution
60.362858
r1_65k_T3_it006_model.star:_rlnCurrentResolution
52.817500
r1_65k_T3_it007_model.star:_rlnCurrentResolution
46.948889
r1_65k_T3_it008_model.star:_rlnCurrentResolution
46.948889
r1_65k_T3_it009_model.star:_rlnCurrentResolution
42.254000
r1_65k_T3_it010_model.star:_rlnCurrentResolution
42.254000
r1_65k_T3_it011_model.star:_rlnCurrentResolution
38.412728
r1_65k_T3_it012_model.star:_rlnCurrentResolution
35.211667
r1_65k_T3_it013_model.star:_rlnCurrentResolution
30.181429
r1_65k_T3_it014_model.star:_rlnCurrentResolution
24.855294
r1_65k_T3_it015_model.star:_rlnCurrentResolution
24.855294
r1_65k_T3_it016_model.star:_rlnCurrentResolution
22.238948
r1_65k_T3_it017_model.star:_rlnCurrentResolution
21.127000
r1_65k_T3_it018_model.star:_rlnCurrentResolution
21.127000
r1_65k_T3_it019_model.star:_rlnCurrentResolution
20.120953
r1_65k_T3_it020_model.star:_rlnCurrentResolution
20.120953
r1_65k_T3_it021_model.star:_rlnCurrentResolution
19.206364
r1_65k_T3_it022_model.star:_rlnCurrentResolution
19.206364
r1_65k_T3_it023_model.star:_rlnCurrentResolution
19.206364
r1_65k_T3_it024_model.star:_rlnCurrentResolution
19.206364
r1_65k_T3_it025_model.star:_rlnCurrentResolution
17.605833
r1_65k_T3_it000_model.star:_rlnLogLikelihood
0.000000
r1_65k_T3_it001_model.star:_rlnLogLikelihood
2.419361e+08
r1_65k_T3_it002_model.star:_rlnLogLikelihood
1.287468e+08
r1_65k_T3_it003_model.star:_rlnLogLikelihood
1.288204e+08
r1_65k_T3_it004_model.star:_rlnLogLikelihood
1.290268e+08
r1_65k_T3_it005_model.star:_rlnLogLikelihood
1.514316e+08
r1_65k_T3_it006_model.star:_rlnLogLikelihood
1.746277e+08
r1_65k_T3_it007_model.star:_rlnLogLikelihood
1.749215e+08
r1_65k_T3_it008_model.star:_rlnLogLikelihood
2.232392e+08
r1_65k_T3_it009_model.star:_rlnLogLikelihood
2.233649e+08
r1_65k_T3_it010_model.star:_rlnLogLikelihood
2.476562e+08
r1_65k_T3_it011_model.star:_rlnLogLikelihood
2.477238e+08
r1_65k_T3_it012_model.star:_rlnLogLikelihood
2.794522e+08
r1_65k_T3_it013_model.star:_rlnLogLikelihood
3.110899e+08
r1_65k_T3_it014_model.star:_rlnLogLikelihood
3.775772e+08
r1_65k_T3_it015_model.star:_rlnLogLikelihood
4.961340e+08
r1_65k_T3_it016_model.star:_rlnLogLikelihood
4.961726e+08
r1_65k_T3_it017_model.star:_rlnLogLikelihood
5.862665e+08
r1_65k_T3_it018_model.star:_rlnLogLikelihood
6.383546e+08
r1_65k_T3_it019_model.star:_rlnLogLikelihood
6.383787e+08
r1_65k_T3_it020_model.star:_rlnLogLikelihood
6.891441e+08
r1_65k_T3_it021_model.star:_rlnLogLikelihood
6.892013e+08
r1_65k_T3_it022_model.star:_rlnLogLikelihood
7.396992e+08
r1_65k_T3_it023_model.star:_rlnLogLikelihood
7.397123e+08
r1_65k_T3_it024_model.star:_rlnLogLikelihood
7.397523e+08
r1_65k_T3_it025_model.star:_rlnLogLikelihood
7.397231e+08
How do it find the convergence of the classification (here, the
loglikelihood is keep increasing during the iterations)?
Also , I start to notice noise pattern in the background of 25th
iteration class averages as compared to the 14th iteration where it is
uniform.
I got the similar results with regularization parameter T = 2 or 3.
Please help to understand this.
thanking you
with kind regards
Mani.
|