Long-read genome sequencing and multi-omics in aging and neurodegeneration.

Structural variants (SVs) are a major source of genetic variation yet remain underexplored in healthy aging and neurodegenerative diseases. We performed nanopore long-read genome sequencing (lrGS) on 551 deeply-phenotyped individuals from Stanford's Aging and Memory Study and Alzheimer's Disease Research Center, generating a comprehensive SV map integrated with matched methylation, transcriptomic, and proteomic data. Over 60% of SVs identified by lrGS were not detected with short-read WGS, including many poorly tagged by single-nucleotide variants (SNVs). We discovered >60,000 SV-QTLs across molecular traits and showed that SVs were more likely than SNVs to be fine-mapped as causal. Colocalization with Alzheimer's and Parkinson's disease GWAS implicated SVs at multiple loci, including TMEM106B, BIN3, and NBEAL1. Multi-omic outlier enrichment and Bayesian modeling prioritized rare functional SVs near known risk genes. Combined, these data reveal widespread regulatory SVs in healthy aging and neurodegeneration, underscoring the importance of lrGS in deciphering complex genetic architecture.