Psychiatric disorders such as autism spectrum disorder (ASD) affect millions of individuals and their families worldwide. Genetic risk plays an important role in the etiology of ASD, and we have shown that changes in brain mRNA expression at the isoform-level, rather than overall changes in gene expression, show the largest effect sizes and genetic enrichments. However, the extent of isoform diversity in brain, and its dysregulation in disorders such as ASD, is vastly underexplored because many isoforms cannot be resolved at the cell type-level by commonly used short-read RNAseq technologies and are yet to be directly profiled. Thus, there is an urgent need to assess the role of isoform expression and its relation to ASD-associated genetic variation − with precise cell-type and spatial specificity to understand the neurobiological mechanisms through which it confers disease risk. Our goal is to address these and other deficits in our understanding of the landscape of isoform expression in neurotypical and diseased brain. As such, we will use long-read RNA isoform sequencing to generate isoform maps and characterize the structure, expression abundance and usage of full-length isoforms in 3 brain regions implicated in ASD. We will further complement by deeply profiling 50 high-confidence ASD risk genes with evidence of isoform dysregulation, to identify isoform expression patterns at the cellular level. In Aim 1 we will perform whole-transcriptome full-length (FL) RNA isoform sequencing (IsoSeq) of postmortem hippocampus (HC) and striatum (STR) brain tissues of ASD cases and neurotypical controls, and combine the data with our existing data for prefrontal cortex (PFC) to construct a comprehensive map of isoform expression across the 3 brain regions. Notably, we will also use our reference maps of normal and dysregulated full-length isoform expression as priors for analyses of a compendium of >2,000 RNAseq datasets from ASD cases and controls amassed by the PsychENCODE consortium and other efforts, and identify dysregulated isoforms and isoform co-expression network modules. In Aim 2 we will use a complementary approach to profile single-nuclei of the same PFC, HC and STR tissues in ASD and controls to disambiguate isoform expression differences of 50 ASD risk genes at the cell-type level using 10X snIsoSeqCap, which is a novel assay to sequence FL- transcripts of selected genes across their entire length at the single nucleus/cell level. We will perform integrative analyses to identify isoform expression differences between tissues and cell types, and selected isoform changes will be validated by RNA fluorescent in situ hybridization (FISH) and quantitative isoform-specific PCR in selected neuronal and non-neuronal cell types. The maps generated here will improve existing reference genome annotations, and allow to address major outstanding questions regarding isoform expression in human brain. The new data and methodologies will provide a tremendous resource for the neuroscience community, paving the way to better inform neurobiological mechanisms of genetic risk for ASD.

PROJECT NARRATIVE Isoform expression in brain and its dysregulation in autism spectrum disorder (ASD) remains vastly underexplored in part because many isoforms are yet to be directly profiled or annotated, and many cannot be reconstructed based on short-read RNA-seq data alone. Here we will generate a map of full-length, tissue- and cell type-specific, isoform diversity in human brain to understand its impact in transcriptional regulation in ASD, with application to other brain disorders.