What is the correct point to run the “remove duplicate” job?
Is it early on after Topaz run or else where? How do I lessen the number of duplicates picked by topaz?
Hi Ani, you need to adjust the extraction radius in Topaz Extract.
This specifies the minimum allowed distance between adjacent picks, in downsampled pixels (according to the level of downsampling used in the preceding training job).
The default value of 7 will result in many duplicates for most samples, so it needs to be adjusted.
So for example, if training was conducted with 4X downsampling, and the original pixel size was 1Å, an extraction radius of 7 would correspond to a minimum distance of 28Å between picks.
Also keep in mind that if you re-extract ant any point during refinement specifically to utilize the particle shifts from 2D or 3D refinement you may want to run the Remove Duplicates again as some of your particles (hopefully very few or none) may have shifted and converged on the same particle depending on how close your particles are or the nature of the particle you are looking at.