Have you tried smaller subsets of the sequences? If there's alternate splicing involved, just take smaller sections of the sequence and, if you're lucky, it'll be the part that's intact enough to get a hit.

Not that I've tried ever it.