Figure 1

A schematic cartoon for the decomposition of segmental duplications into duplication subunits and the construction of map for putative duplication events. (A) Five hypothetical duplication loci (a-e) are depicted with their duplication history shown below. Note that in real cases the historical duplication directions can only be inferred as duplications occurred in the past and are actually invisible. (B) The segmental duplication data for these five loci are represented by seven pairs of duplicons (boxes connected by dash lines). A total of 202 such pairs exist for 22q11.2 based on sequence comparison. (C) Fifteen duplication subunits (forming six paralogous families) decomposed from the pair-wise alignment information in B. (D) The five duplication loci are grouped and all loci are then aligned to the "a" locus, which is the largest one. Note that the entire locus "a" has to be derived from the merge of left duplicons in SD1 and SD2. 33 such duplication groups were defined for 22q11.2, containing 174 duplication loci (see Figure 2B).