Abstract
Single-nucleus RNA-sequencing (snRNA-seq) has revealed new levels of cellular organization and diversity within the human brain. However, full-length mRNA isoforms are not resolved in typical snRNA-seq analyses using short-read sequencing that cannot capture full-length transcripts. Here we combine standard 10x Genomics short-read snRNA-seq with targeted PacBio long-read snRNA-seq to examine isoforms of genes associated with neurological diseases at the single-cell level from prefrontal cortex samples of diseased and nondiseased human brain, assessing over 165,000 cells. Samples from 25 postmortem donors with Alzheimer's disease (AD), dementia with Lewy bodies (DLB), or Parkinson's disease (PD), along with age-matched controls, were compared. Analysis of the short-read libraries identified shared and distinct gene expression changes across the diseases. The same libraries were then assayed using enrichment probes to target 50 disease-related genes followed by long-read PacBio sequencing, enabling linkage between cell type and isoform expression. Vast mRNA isoform diversity was observed in all 50 targeted genes, even those that were not differentially expressed in the short-read data. We also developed an informatics method for detection of isoform structural differences in novel isoforms versus the reference annotation. These data expand available single-cell datasets of the human prefrontal cortical transcriptome with combined short- and long-read sequencing across AD, DLB, and PD, revealing increased mRNA isoform diversity that may contribute to disease features and could potentially represent therapeutic targets for neurodegenerative diseases.