micrograph selection

[cyanpku]

Hi Catsten,

After the micctfdetermine for 70 micrographs I noticed 5 of them show higher defocus than the range I set (0.5 to 5 um) and the pw of the result didn't match the data. I run the micctfdetermin again on those 5 micrographs only and their defocus are above 6 um. I'd like to exclude them for the segmentation by taking the advantage of the "micrographs selection option". In "sqlite3 spring.db .dump", I found the micrograph id corresponding to the bad micrographs to pass onto the micrographs selection. Here's the error from segment/report.log:

ValueError: You have specified selection criteria, please provide the coordinates with a spring.db from previous segment or segmentrefine3d runs. This spring.db will be used for assessing the criteria.

Do I have to run the segment once with the all the micrographs to get the spring.db for micrograph selection?

Or is there a way to manually update the spring.db of the ctf. values so that I can modified the wrong defocus value determined from the first run with the correct value from the second run? 

Thanks a lot,

Jing

[Carsten Sachse]

Dear Jing,

On Thursday, March 3, 2016 at 2:17:48 PM UTC+1, cyanpku wrote:

Hi Catsten,

After the micctfdetermine for 70 micrographs I noticed 5 of them show higher defocus than the range I set (0.5 to 5 um) and the pw of the result didn't match the data. I run the micctfdetermin again on those 5 micrographs only and their defocus are above 6 um. I'd like to exclude them for the segmentation by taking the advantage of the "micrographs selection option". In "sqlite3 spring.db .dump", I found the micrograph id corresponding to the bad micrographs to pass onto the micrographs selection. Here's the error from segment/report.log:

ValueError: You have specified selection criteria, please provide the coordinates with a spring.db from previous segment or segmentrefine3d runs. This spring.db will be used for assessing the criteria. 

Do I have to run the segment once with the all the micrographs to get the spring.db for micrograph selection?

This error message in principle gives you the hint. Alright, I understand the issue now. As there are many different criteria that one can use for the segmentation, I opted to force use spring.db that came out of a previous SEGMENT or SEGMENTREFINE3D run. Otherwise, the possibilities of selection would be ambiguous. In your particular situation, this is a bit counterintuitive.

 In principle, you have three options now. First, you can manually discard the helices from those particular micrographs in order to get rid of the segments (this way there will actually never be segmented). Second, you could run SEGMENT without those criteria to generate the required spring.db file, followed by another run of SEGMENT applying the relevant micrograph exclusion criteria. Third, there are SQlite database editors, which would allow you to update the CTF parameters of the relevant micrographs. If you are familiar with the SQLALCHEMY interface that comes with SPRING, you can do it with Python or there are also SQlite GUI-based editors.

Best wishes,

Carsten

[cyanpku]

Hi Carsten,

Sorry I spell your name wrong : P

Thanks a lot for the explanation!  I supposed the easiest thing to do is to delete the helix box files corresponding to the bad micrographs. 

Bests,

Jing