Splicing QTL (sQTL) Resources#

Splicing quantitative trait loci (sQTL) identify genetic variants that influence RNA alternative splicing patterns, measured as intron excision ratios or splice junction usage. This resource provides sQTL summary statistics, fine-mapping results, TWAS/qTWAS pre-trained models, and colocalization analyses across brain tissues, blood, and AD-relevant cohorts.

Overview#

sQTL mapping was performed using the FunGen-xQTL pipeline. For bulk tissue, splice junction quantification uses LeafCutter-style intron excision ratios followed by tensorQTL mapping. For single-nucleus data, the ISSAC method quantifies splice site usage via UMI-collapsed junction counts using junctools, followed by binomial GLMM sQTL mapping with metacell aggregation. Fine-mapping was performed using SuSiE-RSS yielding PIPs and credible sets.

Available Datasets#

Brain Tissue — ROSMAP Cohort#

Brain Tissue — Other Cohorts#

Blood / Peripheral Tissue#

Analyses Performed#

Single-Context Fine-Mapping#

Fine-mapping using SuSiE-RSS is applied per splicing phenotype (intron cluster or splice site), providing:

• Posterior inclusion probabilities (PIPs)
• 95% credible sets
• Effect sizes on splice junction usage

ISSAC — Single-Nucleus sQTL Method#

For ROSMAP snuc data, ISSAC implements:

• Metacell aggregation: 23,143 metacells for 7 major cell types and 87,936 metacells for 67 retained subcell types, from 3,177,748 nuclei (530 unique donors from 722 specimens: CUIMC N=424, MIT N=298, 192 shared donors)
• Splice site usage quantification with junctools (UMI-level collapsed)
• Binomial GLMM sQTL mapping with PCG/REML for random effects
• Context-dependent sQTL analyses: AD-biased (FDR < 0.01), sex-biased (FDR < 0.05), and cell-state-dependent

Multi-Context Fine-Mapping#

For ROSMAP multi-region bulk sQTL, mvSuSiE (multivariate SuSiE with MASH prior) integrates splicing signals across DLPFC, PCC, and AC jointly. Multi-context fine-mapping results are at syn69670592.

TWAS / qTWAS Models#

Pre-trained sQTL-based TWAS models for splicing-level AD association analyses:

• TWAS weight models (syn69670600): ROSMAP DLPFC, PCC, AC; MSBB; Knight ADRC
• qTWAS models (syn69670611): quantile regression sQTL models

Colocalization#

ColocBoost multi-context colocalization identifies shared splicing QTL signals across brain regions and cohorts (syn69670597). AD GWAS–sQTL colocalization results are available at syn69865816 and syn69670630.

Data Access#