Genomic Variation in Somatic Human Tissues

Genome variation underlies many differences between humans. Less is known about genomic variation between cells within single individuals. Somatic mosaicism is the occurrence of genetically distinct cells within the same organism. Genomic rearrangements occur in gametes by meiotic recombination and in blood cells by V(D)J recombination that confers diversity in immune response. Mitotic recombination was described in Drosophila Melanogaster as early as 1936 (Stern 1936) and takes part in DNA damage repair (Andersen and Sekelsky 2010). Comparison of cancer genomes to genomes from matched normal tissues from the same subjects has revealed somatic rearrangements in the DNA of the tumor cells (Mullighan et al. 2007; Weir et al. 2007).;Although somatic mosaicism is known to occur, it is frequently assumed that cells from healthy human tissues of the same individual are genetically identical. Consequently, typical research and diagnostic practices analyze DNA from a single tissue (e.g. blood). A better understanding of the scope of human somatic variation would inform whether variation at the somatic level is important in studies of genome variation. This dissertation investigates genome variation in somatic human tissues obtained from several human subjects.;DNA from thirty-two tissues from six individuals was analyzed by array-comparative genomic hybridization (aCGH) for copy-number variation (CNV). Between 3-11 tissues were analyzed per person. Tissue comparisons were performed using DNA from one tissue of each individual as a reference. Some 178 tissue-specific CNVs were identified across all tissue comparisons, of which 73 were validated by secondary methods. Many (70%) of the validated CNVs overlap exonic regions and may affect gene expression. A few (7) of the CNV loci appear as tissue-specific CNVs in the same genomic regions in more than one of the subjects suggesting potential hotspots for somatic genome variation. Somatic CNV breakpoints were found to be enriched near microsatellites suggesting a possible role for the repeats in the somatic CNV formation mechanism. The aCGH ratios for the tissue comparisons suggest somatic mosaicism between cells within the tissues sampled.;Tissues from two individuals were analyzed by whole-genome sequencing for tissue-specific single nucleotide variation (SNV), indels, and structural variation (SV). Two tissues, blood and saliva, were analyzed from one of the subjects. Many candidate tissue-specific variants were called by two sequencing platforms, Illumina and Complete Genomics. Sequence capture and resequencing were performed in attempts to validate variant calls as bona fide tissue-specific variants. However, none of the variations could be validated by orthogonal methods. Tissues from a second individual were also analyzed by whole-genome sequencing to investigate whether the difficulty in identifying bona fide tissue-specific variants was specific to blood and saliva DNA. Many tissue-specific SNVs and indels were called from the tissues of the second subject. Inspection of reads at all of the called loci revealed many errors in genome alignment rather than bona fide tissue-specific variants. Only seven tissue-specific SNV calls were credible after inspection of sequence reads; however, none of them could be validated by pyrosequencing. The failure to identify any bona fide tissue-specific variants by whole-genome sequencing suggests low frequency of somatic variation and demonstrates the challenges of detecting variants in mosaic samples by current sequencing technologies.