Medical Scientist Training Program, Duke University, Durham, North Carolina.
Children's Health and Discovery Initiative, Department of Pediatrics, Duke University, Durham, North Carolina.
Cancer Epidemiol Biomarkers Prev. 2020 Aug;29(8):1606-1614. doi: 10.1158/1055-9965.EPI-20-0113. Epub 2020 May 28.
Genome-wide association studies (GWAS) of childhood cancers remain limited, highlighting the need for novel analytic strategies. We describe a hybrid GWAS and phenome-wide association study (PheWAS) approach to uncover genotype-phenotype relationships and candidate risk loci, applying it to acute lymphoblastic leukemia (ALL).
PheWAS was performed for 12 ALL SNPs identified by prior GWAS and two control SNP-sets using UK Biobank data. PheWAS-traits significantly associated with ALL SNPs compared with control SNPs were assessed for association with ALL risk (959 cases, 2,624 controls) using polygenic score and Mendelian randomization analyses. Trait-associated SNPs were tested for association with ALL risk in single-SNP analyses, with replication in an independent case-control dataset (1,618 cases, 9,409 controls).
Platelet count was the trait most enriched for association with known ALL risk loci. A polygenic score for platelet count (223 SNPs) was not associated with ALL risk ( = 0.82) and Mendelian randomization did not suggest a causal relationship. However, twelve platelet count-associated SNPs were nominally associated with ALL risk in COG data and three were replicated in UK data (rs10058074, rs210142, rs2836441).
In our hybrid GWAS-PheWAS approach, we identify pleiotropic genetic variation contributing to ALL risk and platelet count. Three SNPs known to influence platelet count were reproducibly associated with ALL risk, implicating genomic regions containing , proapoptotic protein , and in platelet production and leukemogenesis.
Incorporating PheWAS data into association studies can leverage genetic pleiotropy to identify cancer risk loci, highlighting the utility of our novel approach.
儿童癌症的全基因组关联研究（GWAS）仍然有限，这凸显了需要新的分析策略。我们描述了一种混合 GWAS 和表型全基因组关联研究（PheWAS）方法，以揭示基因型-表型关系和候选风险位点，并将其应用于急性淋巴细胞白血病（ALL）。
使用英国生物库数据，对通过先前的 GWAS 确定的 12 个 ALL SNP 以及两个对照 SNP 集进行了 PheWAS。与对照 SNP 相比，与 ALL SNP 显著相关的 PheWAS 特征被评估为与 ALL 风险（959 例病例，2624 例对照）的关联，使用多基因评分和孟德尔随机化分析。在单 SNP 分析中，对与 ALL 风险相关的特征相关 SNP 进行了测试，在独立的病例对照数据集（1618 例病例，9409 例对照）中进行了复制。
血小板计数是与已知 ALL 风险位点关联最丰富的特征。血小板计数的多基因评分（223 个 SNP）与 ALL 风险无关（ = 0.82），孟德尔随机化也没有表明存在因果关系。然而，COG 数据中 12 个与血小板计数相关的 SNP 与 ALL 风险呈名义相关，其中 3 个在英国数据中得到了复制（rs10058074、rs210142、rs2836441）。
在我们的混合 GWAS-PheWAS 方法中，我们确定了导致 ALL 风险和血小板计数的多效遗传变异。三个已知影响血小板计数的 SNP 与 ALL 风险具有可重复性，这提示了包含 、促凋亡蛋白 和 在内的基因组区域在血小板生成和白血病发生中起作用。
将 PheWAS 数据纳入关联研究可以利用遗传多效性来识别癌症风险位点，凸显了我们新方法的实用性。
