When annotating the unknown sequences, only p-identity is used while the coverage is not considered.