In simple terms yes.
Both the grade and optimize biofeedback scans require frequencies loaded in order to be used. This pretty much means the entire DB is game.
What you have proposed could be done, but you may find that grading again after optimizing may have a significant reduction in benefit to the time invested.
Taking into consideration that 80 frequencies at 3 minutes each is a 4 hour set, this is a good cutoff point that I would consider loading into a remote program loop. So perhaps just grading the multiple HPV sets and take the top 80 or less. Then if you wish, optimize them and then just use them as is. Grading to order the optimized sets would have little effect in the realized efficacy.
For more details, please check the link: