Variant-to-function dissection of the 17q21.31 locus resolves ANKRD1 as a convergent regulatory target
摘要
The 17q21.31 locus is structurally complex, harbouring a common H1/H2 inversion and dense regulatory variation; one of the two SNPs studied lies within the inversion and the other outside it. Disentangling the effects of individual non-coding variants in this context remains challenging. We used reciprocal double-editing in isogenic human iPSCs as a cross-background validation strategy, together with multi-clone replication and a three-way intersection filter (WT-anchored edit and the two reciprocal edit configurations) to resolve variant-linked transcriptional effects for rs77692262 (A > G) and rs79724577 (A > C).
Resultsrs77692262 affected two TGF-β/SMAD-responsive genes, with ANKRD1 replicating across reciprocal and independent contexts, while ACTA2 showed context-dependent effects. By contrast, rs79724577 yielded seven candidates (ANKRD1, CUZD1, GJA5, PRICKLE1, RAB17, RSPO4, SNPH). Independent validation refined these sets: four discovery-set genes, ANKRD1, CUZD1, PRICKLE1, RAB17, replicated for rs79724577 across edited clones/reciprocal contexts, defining a compact, reproducible signature; ANKRD1 also replicated for rs77692262 and was recovered as the sole consistent DEG when rs77692262 was retested in the reciprocal configuration, making it the most clone-invariant target across both variants. GJA5 appeared in discovery but did not replicate and is treated as hypothesis-generating.
Motif analyses were consistent with disrupted T-box/SMAD input at rs77692262 and loss of RFX-family recognition at rs79724577. Long-read CpG methylation profiling detected no significant differences at promoters/gene bodies or at/near the SNP sites, supporting a methylation-independent mechanism involving altered transcription-factor occupancy.
ConclusionsThese data identify reproducible, variant-linked transcriptional effects at 17q21.31 within a fixed H1/H1 haplotypic background, centered on ANKRD1 and a validated four-gene signature for rs79724577, and outline a generalizable, replication-first strategy for variant-to-function dissection in structurally polymorphic regions.