Isoform-level transcriptome-wide association uncovers extensive novel genetic risk mechanisms for neuropsychiatric disorders in the human brain

Abstract

Integrative methods, like colocalization and transcriptome-wide association studies (TWAS), identify transcriptomic mechanisms at only a small fraction of trait-associated genetic loci from genome-wide association studies (GWAS). Here, we show that a reliance on reference functional genomics panels of only total gene expression greatly contributes to this reduced discovery. This is particularly relevant for neuropsychiatric traits, as the brain expresses extensive, complex, and unique alternative splicing patterns giving rise to multiple genetically-regulated transcript-isoforms per gene. We introduce isoTWAS, a scalable, multivariate framework to integrate genetics, isoform-level expression, and phenotypic associations in a step-wise testing framework. Multivariate predictive models were trained using isoform-level expression across tissues from the Genotype-Tissue Expression Project and in the developing and adult human brain from PsychENCODE. Across five neuropsychiatric traits, isoTWAS dramatically increased discovery of trait associations within GWAS loci, capturing 92 unique loci compared with 27 using gene-level TWAS. These power gains reflected a ~2-fold increase in the number of testable genes, an ~15-35% increase in total gene expression prediction accuracy, and the ability to jointly capture expression and splicing mechanisms. Results from extensive simulations showed no increase in false discovery rate and reinforce isoTWAS's advantages in prediction and trait mapping power over TWAS, especially when genetic effects on expression vary across isoforms of the same gene. We illustrate multiple biologically-relevant isoTWAS-identified trait associations undetectable by gene-level methods, including isoforms of AKT3, GIGFY2, and KMT5A with schizophrenia risk. The isoTWAS framework addresses an unmet need to consider the transcriptome on the transcript-isoform level to maximize discovery of trait associations, especially for brain-relevant traits.

Cross-ancestry, cell-type-informed atlas of gene, isoform, and splicing regulation in the developing human brain

“…We next conducted an isoform-centric transcriptome-wide association study (isoTWAS) ( 65 ), to identify genes and their isoforms whose cis regulated expression is associated with SCZ risk. isoTWAS identified 536 isoforms across 271 unique genes with significant isoTWAS associations (Bonferroni-corrected P < 0.05 across genes, FWER-corrected P < 0.05 across transcripts of the same gene, permutation P < 0.05; Methods).…”

Genomics yields biological and phenotypic insights into bipolar disorder

“…Bulk brain-derived eQTL data from PsychENCODE 19 and BrainMeta 20 were integrated with the GWAS results using TWAS as implemented in FUSION 21 and isoTWAS 22 as well as summary-data-based Mendelian randomization (SMR) with subsequent HEIDI (heterogeneity in dependent instruments) tests 23 . All analyses were performed for the multi-ancestry GWAS summary statistics, and for each ancestry group separately, both with and without self-report data where possible (Supplementary Tables 34-40).…”

“…We also applied the recently developed isoform-level TWAS method (isoTWAS) 22 , using precomputed weights per isoform derived from adult PsychENCODE data, as provided by the isoTWAS developers (https://zenodo.org/record/6795947#.Y8mi2-zMLBI). In total we tested the 7,530 genes, each with varying numbers of isoforms, that had positive heritability with a p-value < 0.05 within the PsychENCODE data.…”

Multi‐Omics Studies in Historically Excluded Populations: The Road to Equity

“…Recently, Bhattacharya et al 93 developed a new method: isoformlevel TWAS (isoTWAS). This method performed multivariate predictive modeling using isoform level expression instead of total gene expression across tissues from GTEx and PsychENCODE dataset.…”