I suggest isolating this "scratchpad" tree from the real tree.
One of the cornerstones of Geni is "one person, one profile". The approach here makes a hash of my own tree by making two Nikettis.
The Niketti of legend was the daughter of Opechancanough "Mangopeesomon", paramount chief of the Powhatan and presumably of "Cleopatra", of the Powhatan. The dates are just barely possible but I wouldn't blame anyone for doubting the whole thing.
But the new Nicketti Hughes, Person of legend has been made the daughter of Necotowance, King of the Wyanoak, and so attached to the main tree from which relationships can be calculated. That's a potential source of confusion.
Another problem is that Geni has traditionally disfavored profiles that represent abstract concepts, such as DNA nodes, multiple generations, academic successions, and so on.
A better way to achieve the same objective might be to merge the two versions of Niketti, and start a Niketti DNA project centered around identifying her supposed descendants.
Then, and only then, proceed to evaluate the evidence for Trader Hughes, and for Niketti herself, making firm changes to the tree as you proceed.
I would not be sorry to lose this line of descent, but I would want it done by standard genealogical principles, after an open period of discussion.