Support protein sequences

[fedarko]

Some overlap with #12, maybe.

There are lots of examples of creating dot plots of amino acid sequences (and I think the original Gibbs/McIntyre 1970 paper also created these dot plots?). It would be nice to support this here.

The obvious change is that it isn't immediately clear what the "complement" of an amino acid / stop codon is -- so, in these cases, I think it makes sense to just only include "forward" matches in the dot plot matrix.

A nice side effect of this is that, if we don't care about complementing, then there's no need for us to impose restrictions on the input characters -- so it may make sense to market this as a "mode" which could also be used for comparing arbitrary strings. (Of course, we should still probably check for weird characters that can mess with strings, I guess...?)

But it would be easier for the short term to just strictly limit our focus to protein sequences. As with #21, maybe add a protein parameter to matrix initialization that just (1) changes the allowed alphabet and (2) prevents reverse-complementary / palindromic match detection.