I am wondering what is the definition of the “particle diameter” in template picking. If I have a rod-like template, which dimension should I use for diameter? If I have two templates with different sizes, will same diameter used for two templates?
The particle diameter in template picking is used for two things:
It sets the radius of the circular window that is applied to incoming templates, to remove any signal from the corners/crowding.
It is used to define the minimum particle distance allowed between picks. There is another parameter (min. separation dist) that is a fractional number, that is multiplied with the particle diameter to set this minimum distance.
In your case with a rod-like template, you should set the particle diameter to a value slightly larger than the length of the rod. This makes sure the circular mask will not clip the ends of the rod (in side view). Then you should set the min separation distance to a small value (say 0.2) depending on the aspect ratio of the rod. This makes sure you don’t lose rods that are close together.
Finally, I would recommend doing 2D classification on your picked particles, then do select-2D twice, one selecting side views and once selecting top views. Then, do 2D classificiation in each subset again, but use different “circular mask diameters” for those 2D class runs, based on the different diameter of the side and top view. Then you can combine the good particles from both views into abinito reconstruction and reifnement.