percentidentity

Percent identity is a metric used in sequence analysis to quantify how similar two sequences are at the residue level after an alignment. It is commonly defined as the percentage of positions in the aligned region that contain identical symbols in the two sequences.

Calculation and interpretation can vary slightly by tool. Typically, percent identity is calculated as the number of identical matches divided by the length of the alignment, multiplied by 100. In many implementations, gaps are counted as mismatches unless the method specifies otherwise or reports identity excluding gaps. Some tools also report identity only over the residues that are aligned, ignoring unaligned overhangs. Percent identity is often reported together with alignment length and may be complemented by percent similarity, which also counts conservative substitutions.

Percent identity is used to infer relatedness and homology, to cluster sequences, and to assess the quality of alignments. Higher identity generally indicates closer similarity or common ancestry, while lower identity suggests greater divergence. However, identity alone does not capture functional similarity, as substitutions can be conservative or context-dependent, and alignment choices, scoring matrices, and gap penalties influence the result.

Limitations include dependence on the alignment method, region selection, and sequence length. Short alignments tend to yield inflated identity values, and highly repetitive or low-complexity regions can bias results. Therefore, percent identity is most informative when interpreted alongside alignment coverage, the overall score, and other metrics such as similarity scores, e-values, and bit scores. Related concepts include percent similarity and sequence coverage.