Hi,

This is odd.

Gain correction is done at the very beginning of the program,
regardless of applying local fitting or not.

Are you sure the gain reference is the same for all of the movies?
Do trajectories of these failed movies look sensible?
If you run motion correction with 1 x 1 patch (= no local fitting),
do these 110 movies have streaks and others not?

Best regards,

Takanori Nakane

On 2018/10/06 21:14, Chen Sun wrote:
> Hi Takanori,
> 
> 
> Thank you for your reply. However, for those that failed to fit local motion model, it seems that they are not gain corrected. Except for the 110 movies that were reported with this error, the output micrographs of other movies don't have streaks. And this problem is reproducible. I just confirmed it by rerun the relion job with the same parameters that I included in the first email.
> 
> 
> Here is one of the output .mrc file with the error "too few valid local trajectories to fit local motion model":
> 
> 
> [cid:deb539c2-74f5-4c95-a508-68739ecb12bd]
> 
> 
> Thanks,
> 
> Chen
> 
> 
> ________________________________
> From: Takanori Nakane <[log in to unmask]>
> Sent: Saturday, October 6, 2018 2:24:56 AM
> To: Chen Sun; [log in to unmask]
> Subject: Re: [ccpem] Relion/3.0 motion correction
> 
> Hi,
> 
>   >  "too few valid local trajectories to fit local motion model"
>   >  Does it indicate the bad movie quality? Or is it caused by wrong
> settings?
> 
> This means that many patches were rejected because trajectories
> did not converge. So only the global trajectory was applied.
> You might be able to fix this by reducing the number of patches,
> grouping frames and/or increasing the B factor. That being said,
> I would not worry about this. I do not spend time optimising
> parameters for initial motion correction. As long as you run Bayesian
> Polishing afterwards, initial trajectories are not very important (see
> our preprint).
> 
> Best regards,
> 
> Takanori Nakane
> 
> On 2018/10/06 0:33, Chen Sun wrote:
>> Dear Relion users and developers,
>>
>>
>>
>> I have two questions regarding motion correction with relion/3.0-beta-sm52.
>>
>>
>>
>>     1.  I have been running motion correction with relion's own implementation but it skipped some of the movies and reported the following error message:
>>
>>
>>                         "too few valid local trajectories to fit local motion model"
>>
>>
>>
>>              Does it indicate the bad movie quality? Or is it caused by wrong settings?
>>
>>
>>              Here is the printed command:
>>
>>
>> `which relion_run_motioncorr` --i Import/job001/movies.star --o MotionCorr/job018/ --first_frame_sum 1 --last_frame_sum 0 --use_own  --j 1 --bin_factor 1 --bfactor 150 --angpix 0.9 --voltage 300 --dose_per_frame 1.2 --preexposure 0 --patch_x 5 --patch_y 5 --gainref micrographs/Movies/Gain.mrc --gain_rot 0 --gain_flip 0 --dose_weighting
>>
>>
>>
>>
>>        2. When I run motioncor2 with relion's GUI, some of the movies were summed and corrected into totally black images. The other movies are fine.
>>
>>
>>             Here are my settings for motioncor2:
>>
>>
>>              `which relion_run_motioncorr` --i Import/job001/movies.star --o MotionCorr/job018/ --first_frame_sum 1 --last_frame_sum 0 --use_motioncor2  --motioncor2_exe /apps/motioncor2/1.1.0/motioncor2 --gpu "2" --bin_factor 1 --bfactor 150 --angpix 0.9 --voltage 300 --dose_per_frame 1.2 --preexposure 0 --patch_x 5 --patch_y 5 --gainref micrographs/Movies/Gain.mrc --gain_rot 0 --gain_flip 0 --dose_weighting
>>
>>
>> Has anyone also faced these problems before? And how did you work around them? Any help would be appreciated.
>>
>>
>> Best,
>> Chen
>>
>>
>> ########################################################################
>>
>> To unsubscribe from the CCPEM list, click the following link:
>> https://www.jiscmail.ac.uk/cgi-bin/webadmin?SUBED1=CCPEM&A=1
>>
> 
> 
> ########################################################################
> 
> To unsubscribe from the CCPEM list, click the following link:
> https://www.jiscmail.ac.uk/cgi-bin/webadmin?SUBED1=CCPEM&A=1
> 

########################################################################

To unsubscribe from the CCPEM list, click the following link:
https://www.jiscmail.ac.uk/cgi-bin/webadmin?SUBED1=CCPEM&A=1