number of particles belong to each model in segmultirefine3D

cyanpku

unread,

Mar 18, 2016, 9:29:06 AM3/18/16

to emspring

Dear Carsten,

I guess during the segmultirefine3d, volumes with different helical symmetry compete for segments. Do you know where I can find out how many segments go to each volume? Some of the symmetry models become empty volume after ~ 3 iterations, does it mean they are the wrong symmetry? Another puzzling observation is that some of the symmetry models show inverted contrast (protein become black). The structure seems more or less correct if the contrast is flipped back. Is it another indication of wrong helical symmetry? Any comment will be greatly appreciated!

Jing

Carsten Sachse

unread,

Mar 18, 2016, 6:05:24 PM3/18/16

to emspring

Dear Jing,

On Friday, March 18, 2016 at 2:29:06 PM UTC+1, cyanpku wrote:

Dear Carsten,
I guess during the segmultirefine3d, volumes with different helical symmetry compete for segments.

This assumption is correct.

Do you know where I can find out how many segments go to each volume?

There will be a diagnostic file generated for each model, which contains the respective statistics.

Some of the symmetry models become empty volume after ~ 3 iterations, does it mean they are the wrong symmetry?

In case they are really empty, this suggests that those symmetries are not present in the data set.

Another puzzling observation is that some of the symmetry models show inverted contrast (protein become black).

I assume this could come from poor centering of the segments. Again this will be visible in the statistics diagnostic output.

I hope this helps.

Best wishes,

Carsten

cyanpku

unread,

Mar 19, 2016, 8:12:04 AM3/19/16

to emspring

Hi Carsten,
Thanks for your explanation. Just to make sure the segment numbers for each model is the "asymetric unit before"?
And in comparison of the model with the correct contrast, models with inverted contrast indeed show larger x shift up to the x range limit, I'll further restrain it.
Cheers,
Jing

Carsten Sachse

unread,

Mar 19, 2016, 8:18:08 AM3/19/16

to emspring

Dear Jing,

On Saturday, March 19, 2016 at 1:12:04 PM UTC+1, cyanpku wrote:

Hi Carsten,
Thanks for your explanation. Just to make sure the segment numbers for each model is the "asymetric unit before"?

Yes.This would be the appropriate number asymmetric unit before symmetrization and after.

And in comparison of the model with the correct contrast, models with inverted contrast indeed show larger x shift up to the x range limit, I'll further restrain it.

If you're certain that you have constant diameters of helices, it is the right thing to do.