对耳硬化症位点的证据校准再分析产生了一个可检验的髓系状态假说。

文献信息

中文摘要

背景： 耳硬化症是耳囊骨重塑的局灶性疾病，但变异至细胞状态的机制仍未阐明。方法：我们对耳硬化症GWAS、破骨细胞cis-eQTL、表观基因组、合并Micro-C及人类单细胞RNA测序数据集进行了完全二次性的假说生成分析。结果：在一个因跨层可处理性而非前瞻性全基因组优越性而被选中的、数据丰富的13号染色体位点，共定位在传统先验下与共享区域信号相容，但强烈依赖先验，且未能将COL4A2-AS2、COL4A2或RAB20确定为唯一效应转录本。七个可信变异与一个成熟破骨细胞ATAC元件重叠，合并的破骨细胞环状连接将风险/COL4A2-AS2区间与RAB20相连；相关PBMC接触及缺乏供体解析矩阵排除了破骨细胞特异性调控主张。在一个缺乏chr13基因型的独立12供体骨转移数据集中，RAB20标记了一个与AP1/CTNNB1相关的髓系状态，而一个四供体非肿瘤股骨头数据集仅提供了CXCL8-CXCR2候选轴的跨情境方向性优先排序。对所有27个已报道GWAS位点的简化比较显示，chr13位点属于在成熟破骨细胞开放染色质中具有可信变异的位点之一，但在所检查的注释中并非唯一最强。结论：没有分析数据集同时测量风险单倍型与RAB20相关状态，也没有原始数据集采样耳硬化症病灶。因此，遗传/拓扑学与转录组学观察仍是独立的证据块，产生的是一个可证伪的髓系状态假说，而非确立机制或因果通路。

英文摘要

Background: Otosclerosis is a focal disorder of otic capsule bone remodeling, but variant-to-cell-state mechanisms remain unresolved. Methods: We performed an entirely secondary hypothesis-generating analysis of otosclerosis GWAS, osteoclast cis-eQTL, epigenomic, pooled Micro-C, and human single-cell RNA-sequencing datasets. Results: At a data-rich chromosome 13 locus selected for cross-layer tractability rather than prospective genome-wide superiority, colocalization was compatible with a shared regional signal under the conventional prior but was strongly prior-sensitive and did not resolve COL4A2-AS2, COL4A2, or RAB20 as a unique effector transcript. Seven credible variants overlapped one mature osteoclast ATAC element, and merged osteoclast loop calls connected the risk/COL4A2-AS2 interval with RAB20; related PBMC contacts and the absence of donor-resolved matrices precluded an osteoclast-specific regulatory claim. In a separate 12-donor bone metastasis dataset lacking chr13 genotypes, RAB20 marked an AP1/CTNNB1-associated myeloid state, while a four-donor non-tumor femoral head dataset provided only directional cross-context prioritization of a CXCL8-CXCR2 candidate axis. A simplified comparison across all 27 reported GWAS loci showed that the chr13 locus was among loci with credible variants in mature osteoclast open chromatin but was not uniquely strongest across the examined annotations. Conclusions: No analyzed dataset jointly measured the risk haplotype and the RAB20-associated state, and no raw dataset sampled an otosclerosis lesion. Thus, the genetic/topological and transcriptomic observations remain separate evidence blocks that generate a falsifiable myeloid state hypothesis rather than establish a mechanistic or causal pathway.