Genome-wide detection and analysis of segmental duplications

Segmental duplications play fundamental roles in both genomic disease and gene evolution. To understand their organization and impact within the human genome, I have developed the computational tools necessary to detect and analyze large duplication events. Recent segmental duplications (≥1kb and ≥90% identity) represent over 5% of assembled human genome and the high frequency of these duplications confounds genome assembly---as duplications and true overlaps are often indistinguishable. To overcome this, I used Celera whole-genome shotgun sequence reads to test all public sequences for duplications, creating a database of segmental duplication. This database will greatly aid the assembly of the human genome as well as the study of duplications. Using this database combined with whole genome BLAST analysis, I have studied the pattern and nature of duplications both on a genome scale and at the chromosome level (specifically chromosome 22). Relative to other sequenced organisms, the human genome is composed of a large percentage of highly similar duplications, which appear to have arisen continuously over the past 40 million years. The pattern of duplication varies greatly between chromosomes. Duplications often show enrichment in pericentromeric and subtelomeric regions, consisting of juxtaposed mosaic sequence blocks that are found on multiple chromosomes. This pattern supports the model of preferential and relatively exclusive duplication between such regions after initial seeding events. Detailed comparative FISH studies reveal that the pericentromeric region of chromosome 22 is particularly evolutionary unstable. It varies greatly within the great apes, and containing a human specific duplication event of at least 600 kb. Intrachromosomal duplications also demonstrate interspersed clustering unlike traditional views of tandemly repeated clusters. Such regions often show a mosaic pattern of duplicated modules suggesting a complex history of transposition created by predisposition of duplicated material to create further duplications and rearrangements. Through detailed analyses of chromosome 22, I have identified the creation or modification of 11 transcripts. Whole genome analyses show that duplicated genes on average are less conserved when compared to mouse than genes in unique regions. Together these studies suggest segmental duplications have been an ongoing process of primate evolution contributing to gene evolution and to the rapid remodeling and transformation of genome architecture among closely related species.