Ab inito particle limit

[jpd]

Hi,
ab inito, asking for 1 class, has repeatedly stopped early.
with ~1M particles input, only 262,800 are used in the abinitio.
we are not happy with our output and would like to force it to
use more/all of the particles.

we have seen related threads (ie)

3D ab initio job does not load entire particle stack

For background, at the end of processing I realized I needed to expand my box size to include density for my entire molecule. I attempted to simply re-extract the whole particle set, but got an error about needing to re-extract separately if I had different alignment pixel sizes (they all should have been the same as they came from the same micrographs). The error: “AssertionError: Particles must all have the same alignment pixel size. If multiple particle stacks were input into the job, please…

but have not gotten around this.

we have tried entering the “number of particles to use in optimization”
but this is ignored. We would like to get some control of this parameter.

thanks
John Donohue

[olibclarke]

Don’t have a direct solution for you re the original question, but are you 100% sure you need 1M particles for a single class ab initio?

Modifying parameters, cleaning up the stack first or using multiple classes in ab initio may be faster ways to get a better initial model.

[Mark-A-Nakasone]

If you must force all particles in a single Ab-initio class try increasing Maximum resolution (Angstroms) to something high and also Initial resolution could go up. For 1M particles you are better off using multi class or really cleaning them up. 3DVA, 3D Class, etc are all great but probably should clean up with other jobs before this.

I think 262K are going into a class for a reason. What happens if you take the input of 1M particles for Ab-initio with 4-6 classes (maybe set class similarity to 0.8 if you think they are very similar) and send it to heterogeneous refinement ? Would the volumes be the same in the end or will these jobs filter out some junk ?

It could be worth putting the micrographs and particles together in the curate exposure job before any of this to see if the 1M are coming from good micrographs.