Dear Mike,
thanks for the hint. I normalized the map between 0.000998 and 0.999 in
a similar way and it seems to work. Thanks a lot!
Lubo
On 06/23/2014 10:00 PM, Mike Strauss wrote:
> Hi Lubomir,
>
> I’ve had a similar problem. Try adding 0.000001 to the mask and then multiply it by 0.99998. In my case I had a couple pixels just over 1.0 for some reason, and this seemed to fix it.
>
> Good luck.
>
> mike
>
> On Jun 24, 2014, at 5:40 AM, Lubomir Kovacik <[log in to unmask]> wrote:
>
>> Dear dr. Scheres,
>>
>> I am trying to run a 3D refinement with a reference mask created by the EMAN's e2proc3d.py programme, but I get this Relion error:
>>
>> MlOptimiser::solventFlatten: ERROR solvent mask should contain values between 0 and 1 only...
>> File: src/ml_optimiser.cpp line: 2147
>>
>> The error occurs in the just before the 2nd iteration when the reference map is set to be on absolute grayscale.
>> I have checked both the values and the header in the mrc of the mask (in Matlab), both seem to be ok:
>>
>> Minimum map density 0.0000
>> Maximum map density 1.0000
>> MRC mode 2 (32-bit real)
>>
>> The relion command was this:
>>
>> mpirun -n 4 `which relion_refine_mpi` --o Refine3D/run19 --auto_refine --split_random_halves --i run18_data-1000ptcs.star --particle_diameter 220 --angpix 1.8015 --ref IM/run18_class001.mrc --ini_high 40 --ctf --ctf_corrected_ref --flatten_solvent --solvent_mask IM/run18_class001_emanmaskadj.mrc --oversampling 1 --healpix_order 1 --auto_local_healpix_order 4 --offset_range 5 --offset_step 2 --sym C1 --low_resol_join_halves 40 --norm --scale --j 2 --memory_per_thread 1 &
>>
>> However, the error does not appear when I supply the mask created by the relion_postprocess programme. So I believe that the EMAN mrc should be somehow adjusted. Would you know how?
>>
>> I started to get this error in relion-1.2 but upgrade to 1.3 did not solve the problem.
>>
>>
>> With best regards,
>>
>> Lubomir Kovacik
>> --
>>
>> ***********************************************************************
>> Ing. Lubomir Kovacik, Ph.D. tel.: +420 224968000
>> Institute of Cellular Biology and Pathology fax.: +420 224917418
>> 1st Faculty of Medicine, Charles University
>> Albertov 4
>> 12801 Prague
>> Czech Republic
>> ***********************************************************************
>
--
***********************************************************************
Ing. Lubomir Kovacik, Ph.D. tel.: +420 224968000
Institute of Cellular Biology and Pathology fax.: +420 224917418
1st Faculty of Medicine, Charles University
Albertov 4
12801 Prague
Czech Republic
***********************************************************************
|