One of my students asked me about something that I too have noticed before, but I don't really understand it. Occasionally, for a given amino acid sequence, the BLAST scores on DNA Master don't match up correctly with the actual alignment data. This often seems to happen when the alignment includes low-complexity regions (marked by a string of X's in the alignment). The current example we're working with is that DNA Master returns a score of 194 for the alignment between Belthelas 20 and its homolog in BurtonThePup, with two mismatches, both in a low-complexity region. On the other hand, the alignment between Belthelas 20 and its homolog in VitualEligans, which are 100% identical, results in a score of only 186. It seems as though BLAST can lead us to compare our protein of interest to a protein that isn't the most similar one. What's going on here, and how do we deal with it?