Keeping in mind, of course, that this is just a thought exercise (IIUC, the court ruling does not rely on this line of reasoning), you are right that I overlooked the fact that usually "genes do not spring into existence as a result of sex". (Though we still have to consider novel gene mutations regardless of where they come from.)
Since the patent under discussion does not apply to a genome (I doubt any DNA patent does) but rather to particular mutations of the BRCA1 and BRCA2 genes, if the patenter created/discovered the mutation before it's first occurrence "in the wild", then the defense of the original poster would fail. ("First to file" instead of "first to invent" complicates this, but prior art may get us back to the same place.)
That then leaves two potential loopholes (though they do not apply in this case). First, just as pharmaceutical companies can breathe new life into old drug patents by tweaking an "irrelevant" part of the molecule, it is conceivable that companies could patent variations of the mutated BRCA1 and BRCA2 genes that have not yet occurred in the wild. Second, companies could proactively patent gene mutations that have not yet occurred, but that they think will become important (e.g. some other mutation of BRCA1 that they think would cause cancer if it eventually occurred).
I don't know enough about gene mutation rates or interactions to know how feasible these loopholes would be, but I might hesitate to rely on this defense because of these loopholes.