By dividing the database into these categories the highest scoring homologs of each group were identified, regardless of the score of homologs within other categories.