searching for short, nearly identical peptide matches between two genomes

[green tree]

I want to find all of the short, nearly identical peptide matches between two genomes. These matches are ~10bp in length, allowing for 1 or 2 mismatches, but no gaps.

When I do the BLAST search, I can find many such matches. However, I also find longer matches that are below my identity cutoff (e.g. 30bp matches that are only 70% identity).

From what I know about BLAST, it seems possible that there are 10bp perfect matches buried in these 30bp 70% identity matches. Is this correct? If so, can anyone recommend a way to solve this problem, or point me in the right direction?

I will even code this up in python if that is the best way. (I guess I would split my database into words of length 3, then look at every word in my query and calculate the edit distances between all such strings... doesn't sound very fun...)

[green tree]

One option is to get the full alignments (instead of tab-delimited format) and search for identical sub-alignments... doesn't sound very fun...