Hi,
Quite often when processing data from crowded micrographs, particularly of small particles, applying a tight circular mask during 2D classification can help to improve classes and mitigate the influence of neighboring particles.
Currently there is no option to apply a spherical mask during ab initio reconstruction, and we frequently see artifacts from adjacent particles during ab initio in such datasets.
Would it be possible to add an option to apply a spherical mask to the volume at each iteration of ab initio, to deal with such edge cases? There is an option to control windowing the dataset, but this is applied to the particle images before CTF correction, not the volume, and hence is not quite the same thing.
This has been added to heterogeneous refinement, and is useful there - but often initial volume estimation is a critical step for small particles, so anything that could help in difficult edge cases would be useful.
Cheers
Oli
EDIT:
a better example of the issue:
This is using particles recentered after 2D, clean 2D classes from the run above.