Cool. I saw this the other day when you posted it at BioStar (I am Damian Kao at BioStar). Next step is to establish some alignment metrics (some kind of p-value).
There's work I've started on this, actually. It's available in the README of the NtSeq repository[1]. Near the bottom. The probability becomes a mess (or at least, I can't figure out the correct model) when you start including degenerate nucleotides, and I didn't want to release a half-done statistical analysis tool, so I axed it from the current release.
[1] https://github.com/keithwhor/NtSeq