Import into M, one round with options off -> 650A

[joe dobbs]

Hi,

I have a strange issue and wondered if any of you have seen similar:

-refinement in relion of a ribosome reaches 20A

-import into M (using relion_star_downgrade from dynamo2m)

-one iteration with all options turned off to test the import went ok -> 650A empty looking structure (+M crashes)

(as suggested in https://groups.google.com/g/warp-em/c/ZMr-uW8bBzU/m/CweU9wrxCgAJ)

-however, re-extract of the particles from M directly after import and relion_reconstruct -> ribosome is still there

I don't use the pre-rotate particles option, and the star file conversion seems to have gone ok- these potential sources of error are indicated against since I can again reconstruct my structure without realignment after M export.

It is then very strange that one iteration without any options affected the structure in any way- surely it should have just come back with my 20A ribosome, not some 650A thing.

Has anyone seen anything like this?

[Mathew McLaren]

Hi Joe,

That sounds like it could be an incorrect pixel size. If the pixel size you enter for the star file after downgrading is incorrect then it may just be working with noise, although I'm not sure if that would explain why it reextracts with no problems afterwards. The most recent version of Warp/M now supports Relion 3.1 star files and as such you shouldn't need to use the relion_star_downgrade.

Also, does your mask look reasonable? Was it generated in M or imported from something like Relion?

Cheers,

Mat

--
You received this message because you are subscribed to the Google Groups "Warp" group.
To unsubscribe from this group and stop receiving emails from it, send an email to warp-em+u...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/warp-em/3f19ba0a-de1b-46c5-8059-b9802754fe44n%40googlegroups.com.

[joe dobbs]

Hi Mat,

I think possibly wrong pixel size during import is also counter-indicated by the reextraction with no problems, I don't see how it could regenerate properly if the coordinates were off. I'll also note that I also tried to manually make the star file with the same result.

The mask was generated inside M from the unfiltered half-maps and seems reasonable to me.

Thanks for the suggestions.

[Ricardo Righetto]

Some ideas for troubleshooting:

How do the half-maps look like? And the full FSC curve (not just the resolution number)?

Have you tried varying the mask parameters?

Good luck,

--
Ricardo Diogo Righetto

To view this discussion on the web visit https://groups.google.com/d/msgid/warp-em/07fcce3d-f4c5-410e-bf00-bf07505e4cf7n%40googlegroups.com.