The genome of CMV clinical isolates, such as the Merlin strain depicted here (GenBank accession no. NC_006273; ref. 118), consists of long (brown) and short (orange) DNA segments, each of which has unique regions (UL and US) flanked by inverted repeats (TRL/IRL and IRS/TRS). These repeats contain segment-specific sequences (b, b′, c, and c′) as well as a variable number of shared a sequence repeats in direct orientation at the genomic ends and in an inverted orientation at the junction of the two segments. Laboratory-adapted strains often have deletions of multiple genes at the right end of the UL segment and their replacement with genes duplicated from the left end, resulting in longer TRL and IRL regions (dashed boxes) compared with clinical strains (21, 118). The gene names in this region are not always sequential because of historical precedence in nomenclature assignments and because of rearrangements among strains. The relative position and orientation of transcripts corresponding to several genes are shown, along with grouping by their putative functional classifications (24, 31, 43–45, 47). This diagram is a simplification, since some and possibly many of the genes shown here have more than one function and other genes that are not shown likely contribute to the indicated processes.