Which of the following statements about COG is incorrect regarding its features?

• A. Currently, there are 4,873 clusters in the COG databases derived from unicellular organisms
• B. It is constructed by comparing protein sequences encoded in forty-three completely sequenced genomes, which are mainly from prokaryotes, representing thirty major phylogenetic lineages
• C. The interface for sequence searching in the COG database is the COGnitor program, which is based on gapped BLAST
• D. It is a protein family database based on structural classification