[i] Assuming 86% base identity and 89% amino acid identity, this table shows the maximum K sizes and number of chance matches passed to the alignment stage when searching for regions of 100 bases (or 33 amino acids) with at least a 99% chance of detecting the homology. These values reflect our targets for human/mouse alignments. For translated DNA sequences, the F value is multiplied by six to reflect three reading frames on both strands of the query. Even with this multiplication, the specificity for a given sensitivity is several orders of magnitude greater in the amino acid rather than the nucleotide domain.