DNA copy number variations (CNVs) have previously been reported in human cortical neurons from non-diseased patients, but these alterations do not appear to be consistent from cell to cell and appear to be rare among neurons overall. Interestingly, Alzheimer’s disease patients appear to have a higher prevalence of CNVs than non-diseased, although the biological significance of this observation is still largely unknown. Single-cell whole-genome next-generation sequencing holds promise to investigate these variations and the regions in which
they occur in an unbiased manner. Unlike recent advances in single-cell RNA-seq, however, library preparation for single-cell DNA-seq suffers from extremely limited throughput. Furthermore, it is difficult to assess the significance of individual variations from whole-genome sequencing alone, particularly when control samples from non-diseased patients also show some variation at lower frequency. A potential solution is a multi-omics approach, in which information is collected about multiple species of biomolecules simultaneously from each sample, which taken together aid the interpretation of individual observations with respect to biological significance.
This dissertation describes the design and development of a technology to physically separate DNA and RNA and to prepare sequencing libraries from each in parallel from limited starting samples without splitting, which we called Gel-seq. Thirty-two paired DNA and RNA sequencing libraries were successfully prepared from a variety of human and mouse cells lines and from mouse liver tissue using Gel-seq. Sample types could be clearly distinguished from each other based on either genomic copy number or transcriptomic profiles. This dissertation also describes the design and development of a technology to prepare a thousand single-cell whole-genome sequencing libraries in a single run. A proof-of-concept was performed with 87 cells from human and mouse lines. Copy number profiles agreed with bulk, and 96% and 92% of human and mouse cells, respectively, clustered correctly within their cell line based on copy number profile alone. These technologies will help to enable the unbiased characterization of genomic alterations not only in neurodegenerative disorders, but potentially also in other conditions associated with mosaic genomic backgrounds, such as cancer, microbiome disorders, or infectious diseases.