Hi Tofayel,
Have you checked whether there are any "strange" particles, like all-0 or
with very strong pixels or so? This happens more often than one might
suspect, and can cause these problems.
Also, if you add up the particles in all groups from half1 and half1, then
you should get to 7000 again.
HTH,
S
> Hi Sjors and Relion Users,
>
> I am processing a dataset on ribosomes in Relion 1.3 and after sorting
> done at both 2D and 3D classification steps, I am left with around 7000
> particles. I went on to do Refine-3D and relion shows me an error which I
> have pasted below. However, I was going through the archives of ccpem
> mailing list and found out that this question was discussed earlier but
> failed to find out a consensus on the solution. Clearly, relion has some
> problem with some of the groups where particle_sumweight=nan or ZERO !!!
> Looked through previous iterations and all the star files if I had NaN
> assigned to any of the fields but none was found. I wanted to fish out the
> exact micrographs by looking at run*_it*_half1_model.star file and I did
> so. But the micrographs looks fine to me. I also figured out that in
> data_model_groups column, the _rlnGroupNrParticles doesnt add up to 7000
> particles. Clearly, I may be missing out on some important point about how
> grouping in done in relion but my understanding is , if we don't specify
> any group, relion assigns each micrograph to one group and thats how I get
> 164 groups (I started with 164 micrographs) but would also expect values
> in _rlnGroupNrParticles add up to 7000.
>
> Any help would be much appreciated !
>
>
> Auto-refine: Iteration= 19
> Auto-refine: Resolution= 20.5714 (no gain for 1 iter)
> Auto-refine: Changes in angles= 0.572653 degrees; and in offsets=
> 0.477056 pixels (no gain for 0 iter)
> Estimating accuracies in the orientational assignment ...
> 7/ 7 sec
> ............................................................~~(,_,">
> Auto-refine: Estimated accuracy angles= 1.891 degrees; offsets= 1.571
> pixels
> Auto-refine: Angular step= 0.9375 degrees; local searches= true
> Auto-refine: Offset search range= 3.67425 pixels; offset step= 1.22475
> pixels
> CurrentResolution= 20.5714 Angstroms, which requires orientationSampling
> of at least 6.54545 degrees for a particle of diameter 360 Angstroms
> Oversampling= 0 NrHiddenVariableSamplingPoints= 21233664
> OrientationalSampling= 1.875 NrOrientations= 150
> TranslationalSampling= 2.4495 NrTranslations= 9
> =============================
> Oversampling= 1 NrHiddenVariableSamplingPoints= 679477248
> OrientationalSampling= 0.9375 NrOrientations= 1200
> TranslationalSampling= 1.22475 NrTranslations= 36
> =============================
> Estimated memory for expectation step > 0.194003 Gb, available memory =
> 4 Gb.
> Estimated memory for maximization step > 0.168626 Gb, available memory =
> 4 Gb.
> Expectation iteration 19
> 3.02/10.47 min .................~~(,_,"> exp_thisparticle_sumweight= nan
> [oo]
> exp_part_id= 2489exp_iimage=1
> group_id= 140 mymodel.scale_correction[group_id]= 0.974612
> exp_ipass= 0
> sampling.NrDirections(0, true)= 12288 sampling.NrDirections(0, false)= 29
> sampling.NrPsiSamplings(0, true)= 192 sampling.NrPsiSamplings(0, false)=
> 6
> mymodel.sigma2_noise[exp_ipart]=
> 0.00029
> 0.00017
> 0.00011
> 8.5e-05
> 0.00011
> 0.00011
> 0.00011
> 0.0001
> 0.00011
> 0.0001
> 9.5e-05
> 0.00011
> 0.00011
> 0.00011
> 9.4e-05
> 0.00011
> 0.0001
> 9.1e-05
> 8.8e-05
> 7.4e-05
> 6.4e-05
> 5.5e-05
> 4.8e-05
> 3.9e-05
> 3.4e-05
> 3.5e-05
> 3.2e-05
> 3.4e-05
> 3.2e-05
> 3.4e-05
> 3.3e-05
> 3e-05
> 2.8e-05
> 2.5e-05
> 2.4e-05
> 2.2e-05
> 2.1e-05
> 2.1e-05
> 2.3e-05
> 2.2e-05
> 2e-05
> 2e-05
> 1.8e-05
> 1.7e-05
> 1.5e-05
> 1.5e-05
> 1.5e-05
> 1.6e-05
> 1.5e-05
> 1.3e-05
> 1.2e-05
> 1.2e-05
> 1.2e-05
> 1.1e-05
> 1.1e-05
> 1e-05
> 9.7e-06
> 9.2e-06
> 9.6e-06
> 9.2e-06
> 8.5e-06
> 8e-06
> 7.7e-06
> 7.4e-06
> 7.3e-06
> 7e-06
> 6.9e-06
> 6.3e-06
> 6.1e-06
> 6.2e-06
> 5.9e-06
> 5.4e-06
> 5.3e-06
> 5.1e-06
> 5.1e-06
> 4.7e-06
> 4.5e-06
> 4.4e-06
> 4.4e-06
> 4e-06
> 3.9e-06
> 3.9e-06
> 3.7e-06
> 3.5e-06
> 3.4e-06
> 3.3e-06
> 3.3e-06
> 3.1e-06
> 2.9e-06
> 2.9e-06
> 2.7e-06
> 2.6e-06
> 2.5e-06
> 2.5e-06
> 2.4e-06
> 2.3e-06
> 2.3e-06
> 2.1e-06
> 2.1e-06
> 2e-06
> 2e-06
> 2e-06
> 1.9e-06
> 1.9e-06
> 1.7e-06
> 1.7e-06
> 1.7e-06
> 1.5e-06
> 1.6e-06
> 1.5e-06
> 1.4e-06
> 1.4e-06
> 1.4e-06
> 1.3e-06
> 1.3e-06
> 1.3e-06
> 1.3e-06
> 1.2e-06
> 1.2e-06
> 1.1e-06
> 1.1e-06
> 1.1e-06
> 1.1e-06
> 1e-06
> 1e-06
> 9.8e-07
> 9.6e-07
> 9e-07
> 9.2e-07
> 9.1e-07
> 8.8e-07
> 8.5e-07
> 8.2e-07
> 8.3e-07
> 8.1e-07
> 7.8e-07
> 7.6e-07
> 7.3e-07
> 7.4e-07
> 7.1e-07
> 6.9e-07
> 6.8e-07
> 6.7e-07
> 6.6e-07
> 6.4e-07
> 6.3e-07
> 6.2e-07
> 6.1e-07
> 6.1e-07
> 5.8e-07
> 5.8e-07
> 5.9e-07
> 5.8e-07
> 5.6e-07
> 5.5e-07
> 5.5e-07
> 5.3e-07
> 5.2e-07
> 5.2e-07
> 5.2e-07
> 5.1e-07
> 5.1e-07
> 5e-07
> 4.9e-07
> 5e-07
> 4.7e-07
> 4.7e-07
> 4.8e-07
> 4.8e-07
> 4.7e-07
> 4.6e-07
> 4.6e-07
> 4.5e-07
> 4.4e-07
> 4.5e-07
> 4.4e-07
> 4.3e-07
> 4.3e-07
> 4.2e-07
> 4.2e-07
> 3.8e-07
>
> wsum_model.sigma2_noise[exp_ipart]=
> 0.00505
> 0.008294
> 0.008885
> 0.0105
> 0.02199
> 0.02241
> 0.03025
> 0.02976
> 0.03605
> 0.04888
> 0.03513
> 0.05765
> 0.05674
> 0.06473
> 0.06146
> 0.06058
> 0.08316
> 0.07294
> 0.068
> 0.0656
> 0.04934
> 0.05143
> 0.04654
> 0.03894
> 0.03511
> 0.03978
> 0.03836
> 0.03918
> 0.04001
> 0.04065
> 0.04533
> 0.04163
> 0.03614
> 0.03734
> 0.03787
> 0.03303
> 0.03379
> 0.03425
> 0.03926
> 0.03466
> 0.03528
> 0.03365
> 0.03082
> 0.03105
> 0.0288
> 0.03068
> 0.02896
> 0.03348
> 0.03103
> 0.02899
> 0.02711
> 0.02762
> 0.02841
> 0.02494
> 0.02458
> 0.02527
> 0.02241
> 0.02421
> 0.02179
> 0.02525
> 0.02277
> 0.02053
> 0.02133
> 0.01942
> 0.02277
> 0.01946
> 0.02045
> 0.01842
> 0.01772
> 0.02026
> 0.01784
> 0.01644
> 0.0181
> 0.01563
> 0.01669
> 0.0158
> 0.01571
> 0.01547
> 0.01436
> 0.01373
> 0.01335
> 0.01448
> 0.0136
> 0.01273
> 0.01241
> 0.01297
> 0.013
> 0.01181
> 0.01068
> 0.01209
> 0.01102
> 0.01043
> 0.0102
> 0.01051
> 0.00987
> 0.009272
> 0.009781
> 0.009298
> 0.009613
> 0.008501
> 0.008618
> 0.008227
> 0.008696
> 0.008679
> 0.007486
> 0.008003
> 0.008031
> 0.007193
> 0.007495
> 0.007211
> 0.007372
> 0.006475
> 0.006878
> 0.006456
> 0.006751
> 0.006632
> 0.006445
> 0.006353
> 0.00599
> 0.005952
> 0.006162
> 0.006017
> 0.005703
> 0.005698
> 0.005432
> 0.005284
> 0.005236
> 0.005386
> 0.00511
> 0.005097
> 0.004898
> 0.004766
> 0.004801
> 0.004767
> 0.004821
> 0.004455
> 0.00458
> 0.004311
> 0.004356
> 0.004669
> 0.004365
> 0.004334
> 0.004264
> 0.003941
> 0.004314
> 0.004089
> 0.003939
> 0.003849
> 0.003896
> 0.003935
> 0.003769
> 0.003948
> 0.003747
> 0.003863
> 0.003831
> 0.003606
> 0.003647
> 0.003594
> 0.003663
> 0.003462
> 0.003476
> 0.003862
> 0.003499
> 0.003513
> 0.003432
> 0.003445
> 0.003364
> 0.003358
> 0.003457
> 0.003516
> 0.003467
> 0.003608
> 0.003244
> 0.003466
> 0.003389
> 0.003315
> 0.003341
> 0.003351
> 0.003466
> 0.003308
> 0.002911
>
> mymodel.avg_norm_correction= 0.955164
> wsum_model.avg_norm_correction= 677.562
> written out Mweight.spi
> exp_thisparticle_sumweight= nan
> exp_min_diff2[exp_ipart]= 9.9e+100
> slave 3 encountered error: ERROR!!! zero sum of weights....
> File: src/ml_optimiser.cpp line: 3982
> +++ RELION: command line arguments (with defaults for optional ones
> between parantheses) +++
> ====== Continue options =====
> --continue : _optimiser.star file of the
> iteration after which to continue
> ====== General options =====
> --o (OLD_ctX) : Output rootname
> --iter (OLD) : Maximum number of iterations to
> perform
> --tau2_fudge (OLD) : Regularisation parameter (values
> higher than 1 give more weight to the data)
> --flatten_solvent (OLD) : Switch on masking on the references?
> --solvent_mask (OLD) : User-provided mask for the
> references
> --solvent_mask2 (OLD) : User-provided secondary mask
> --tau (OLD) : STAR file with input tau2-spectrum
> (to be kept constant)
> --particle_diameter (OLD) : Diameter of the circular mask that
> will be applied to the experimental images (in Angstroms)
> --join_random_halves (false) : Join previously split random halves
> again (typically to perform a final reconstruction).
> ====== Re-align movie frames =====
> --realign_movie_frames () : Input STAR file with the movie
> frames
> --nr_frames_prior (5) : Number of movie frames to calculate
> running-average priors
> --movie_frames_running_avg (3) : Number of movie frames in each
> running average
> ====== Orientations =====
> --oversampling (OLD) : Adaptive oversampling order to
> speed-up calculations (0=no oversampling, 1=2x, 2=4x, etc)
> --auto_local_healpix_order (OLD) : Minimum healpix order (before
> oversampling) from which auto-refine procedure will use local searches
> --sigma_ang (OLD) : Stddev on all three Euler angles for
> local angular searches (of +/- 3 stddev)
> --sigma_rot (OLD) : Stddev on the first Euler angle for
> local angular searches (of +/- 3 stddev)
> --sigma_tilt (OLD) : Stddev on the first Euler angle for
> local angular searches (of +/- 3 stddev)
> --sigma_psi (OLD) : Stddev on the in-plane angle for
> local angular searches (of +/- 3 stddev)
> --sigma_off (OLD) : Stddev. on the translations
> --skip_align (false) : Skip orientational assignment (only
> classify)?
> --skip_rotate (false) : Skip rotational assignment (only
> translate and classify)?
> --skip_maximize (false) : Skip maximization step (only write
> out data.star file)?
> ====== Corrections =====
> --scale (OLD) : Switch on intensity-scale
> corrections on image groups
> --norm (OLD) : Switch on normalisation-error
> correction
> ====== Computation =====
> --j (1) : Number of threads to run in parallel
> (only useful on multi-core machines)
> --pool (OLD) : Number of images to be processed
> together
> --dont_combine_weights_via_disc (false) : Send the large arrays of
> summed weights through the MPI network, instead of writing large files
> to disc
> --verb (1) : Verbosity (1=normal, 0=silent)
> ====== Expert options =====
> --strict_highres_exp (OLD) : Resolution limit (in Angstrom) to
> restrict probability calculations in the expectation step
> --scratchdir () : Directory (with absolute path, and
> visible to all nodes) for temporary files
> --------------------------------------------------------------------------
> MPI_ABORT was invoked on rank 3 in communicator MPI_COMM_WORLD
> with errorcode 1.
>
> NOTE: invoking MPI_ABORT causes Open MPI to kill all MPI processes.
> You may or may not see output from other processes, depending on
> exactly when Open MPI kills them.
>
>
> Thanking you,
> Tofayel
>
--
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
|