harry potter and the doubly helix

I'd always presumed muggle-born wizards in the Harry Potter books were the result of a genetic mutation. J.K. Rowling is answering all kinds of questions now about the series, and she says I'm wrong:

Katie Mosher: How exactly do muggleborns receive magical ability?

J.K. Rowling: Muggle-borns will have a witch or wizard somewhere on their family tree, in some cases many, many generations back. The gene re-surfaces in some unexpected places.

This is very perplexing, because magic ability otherwise gives all indications of being an autosomal dominant gene (get the gene from one parent, and you have the trait), but now she's talking about it like it is recessive and so can just lurk around unexpressed in your family tree for generations. Harry had a muggle-born mother, so if the gene was recessive, he'd have had only a 50% chance of being a wizard. Same thing with each Hermione and Ron's two children (assuming Ron is really the real father; if, as is at least as likely, Harry is the real father, then the kids would have only a 25% chance of being a wizard.) There are no indications in the book that when people marry muggle-born wizards they consider there to be a 50% chance they'll have a squib kid.

I feel cheated. I feel like the Harry Potter books are no longer scientifically realistic.

Update: Ugh, I've only had this post up a few minutes and already a friend has e-mailed to correct me. I feel like a disgrace to Ravenclaw. Obviously, if the gene is recessive, Harry needs two copies, and so he has the same probability of having wizard spawn as any other wizard, and squibhood is the mutation. (Squibs who breed with a slumming wizard would have a 50% chance of having a wizard child.) Rowling's still not right, though. The muggle-born wizard has to have a wizard somewhere on each side of the family tree.