High-accuracy imputation for HLA class I and II genes based on high-resolution SNP data of population-specific references.
High-accuracy imputation for HLA class I and II genes based on high-resolution SNP data of population-specific references.
复制标题
基于人口特异性参考的高分辨率SNP数据，HLA I和II类基因的高准确性推出。
DOI：
10.1038/tpj.2015.4
复制
发表时间：
2015-12
期刊：
影响因子：
--
通讯作者：
Tokunaga K
中科院分区：
文献类型：
--
作者：
Khor SS;Yang W;Kawashima M;Kamitsuji S;Zheng X;Nishida N;Sawai H;Toyoda H;Miyagawa T;Honda M;Kamatani N;Tokunaga K
经典人类白细胞抗原（HLA）等位基因的统计插补正成为从病例对照全基因组关联研究中精细映射疾病关联信号的不可或缺的工具。然而，目前大多数可用的HLA填补工具是基于欧洲参考人群，不适合直接应用于非欧洲人群。在HLA填补工具中，HIBAG R包是一个灵活的HLA填补工具，配备了广泛的基于人群的分类器;此外，HIBAG R使个人研究人员能够构建自定义分类器。在这里，两个数据集，每个数据集包括来自不同样本量的健康日本人的数据，用于构建自定义分类器。在五个HLA类别（HLA-A、HLA-B、HLA-DRB 1、HLA-DQB 1和HLA-DPB 1）中的HLA填补准确性从使用原始HIBAG参考获得的82.5-98.8%增加到使用我们的定制分类器获得的95.2-99.5%。对于我们的日语分类器，推荐调用阈值（CT）为0.4;相比之下，HIBAG参考文献推荐CT为0.5。最后，我们的分类器可用于识别日本发作性睡病伴卡他汀的风险单倍型，HLA-DRB 1 *15：01和HLA-DQB 1 *06：02，准确率分别为100%和99.7%;因此，这些分类器可用于补充目前广泛可用的全基因组关联研究数据集中缺乏HLA基因分型数据。
Statistical imputation of classical human leukocyte antigen (HLA) alleles is becoming an indispensable tool for fine-mappings of disease association signals from case–control genome-wide association studies. However, most currently available HLA imputation tools are based on European reference populations and are not suitable for direct application to non-European populations. Among the HLA imputation tools, The HIBAG R package is a flexible HLA imputation tool that is equipped with a wide range of population-based classifiers; moreover, HIBAG R enables individual researchers to build custom classifiers. Here, two data sets, each comprising data from healthy Japanese individuals of difference sample sizes, were used to build custom classifiers. HLA imputation accuracy in five HLA classes (HLA-A, HLA-B, HLA-DRB1, HLA-DQB1 and HLA-DPB1) increased from the 82.5–98.8% obtained with the original HIBAG references to 95.2–99.5% with our custom classifiers. A call threshold (CT) of 0.4 is recommended for our Japanese classifiers; in contrast, HIBAG references recommend a CT of 0.5. Finally, our classifiers could be used to identify the risk haplotypes for Japanese narcolepsy with cataplexy, HLA-DRB1*15:01 and HLA-DQB1*06:02, with 100% and 99.7% accuracy, respectively; therefore, these classifiers can be used to supplement the current lack of HLA genotyping data in widely available genome-wide association study data sets.