Integrating single-cell sQTL mapping with deep-learning splicing prediction identifies causal variants under influenza infection.

Publication date: Jul 18, 2026

Realizing the full potential of human genetics requires identifying causal variants and genes underlying association signals. Molecular quantitative trait locus (molQTL) analyses, such as expression QTL (eQTL) and splicing QTL (sQTL), link genetic variants to intermediate molecular phenotypes, but pinpointing causal variants and their regulatory effects remains challenging. Here, we integrate sQTL analysis with deep learning-based splicing effect annotation to identify causal genetic variants and elucidate their functional effects on human phenotypes. Applying a single-cell GWAS method, scHi-HOST, across 96 lymphoblastoid cell lines with and without influenza A virus (IAV) infection, we identify approximately 43,000 sQTL SNP-junctions pairs associated with 217 genes during infection. Integrating sQTL mapping with AI based splice prediction and statistical fine-mapping, we prioritize 57 likely causal variants that affect cis-acting molecular splicing components, including 5′ donor and 3′ acceptor sites. We experimentally validate one such variant, rs2297616, in PARP2, encoding poly (ADP-ribose) polymerase 2. This variant alters the 5′ splice donor site in the second intron of PARP2, generating two protein isoforms differing by 13 amino acids. The derived A allele is associated with the longer protein isoform and increased IAV levels in lymphoblastoid cell lines. CRISPR mediated editing validates the causal effect of this variant on both protein length and IAV infection. Furthermore, these 57 putative causal sQTLs are linked to over 100 GWAS traits, including many variants associated with autoimmune diseases. Our work provides a catalog of causal sQTL with direct splicing impacts, providing causal mechanistic insights from genotype to disease susceptibility.

Open Access PDF

Semantics

Original Article