Discovery and Analysis of Somatic Mutations in Aging Neurons Using Single-Cell DNA Sequencing

Abstract

Single-cell DNA sequencing (scDNAseq) can, in principle, allow us to ask and answer many novel questions in genomics. Approaches so far, however, have suffered from a prohibitive abundance of artifacts derived from the whole-genome amplification (WGA) of the single genome. Here, I offer an improved computational method for analyzing scDNAseq data, Linked-Read Analysis (LiRA). LiRA uses read-backed phasing of candidate somatic single-nucleotide variants (sSNVs) with known germline heterozygous single-nucleotide polymorphisms (gHets) to improve discrimination of artifacts from true mutations. To demonstrate LiRA, I apply it to neurons sampled from the post-mortem brain tissue of individuals of different ages. Advanced age is associated with increased risk of disease in many tissues and organ systems, but the mechanism underlying this phenomenon is poorly understood. While many hypotheses exist and aging may have many underlying molecular causes, one attractive idea, driven by the fact that all cell and tissue types must maintain genomic integrity to properly function, is that the accrual of somatic mutations causes dysfunction that eventually leads to disease. The role of somatic mutations in cancer is well studied, but it is unclear what, if any, role somatic mutations have in aging outside of this context. In analysis of cancer samples, clonal expansion enables bulk sequencing to be used to characterize driver somatic mutations present in tumor and absent from normal tissue. In contrast, somatic mutations driving the aging phenotype could be present in arbitrarily small cell populations or acquired privately in post-mitotic cells, and are undetectable using a similar approach. Using results obtained with LiRA, I describe and analyze the pattern of mutations observed in the aging brain.