NEW YORK (GenomeWeb) – Researchers have analyzed the genetic and phenotypic landscape of the major histocompatibility complex (MHC) region in a Japanese cohort.
The MHC region is one of the most polymorphic regions of the human genome and harbors population-specific variation. It further contains more than 200 genes, including human leukocyte antigen (HLA) genes, which are linked to a number of complex human traits.
Researchers led by Osaka University Graduate School of Medicine's Yukinori Okada performed sequencing-based typing of 33 HLA genes in more than 1,100 individuals of Japanese ancestry to develop a catalog of HLA alleles and used that to both impute genome-wide association study data and perform a phenome-wide association study of more than a hundred clinical phenotypes. As they reported today in Nature Genetics, the researchers found a number of phenotypes, including risk of certain diseases, were linked to HLA genes.
"[O]ur study comprehensively elucidated the genetic and phenotypic landscapes of MHC in the Japanese population," Okada and his colleagues wrote in their paper.
He and his colleagues performed high-resolution typing of 33 HLA-related genes with up to six-digit-level allele information for 1,120 individuals. They used a target-capture technique and sequencing using longer read lengths than earlier efforts. This way, the researchers reported they were able to achieve higher accuracy in classical HLA allele typing than previous next-generation sequencing approaches have.
Of the 33 sequenced HLA genes, nine were classical HLA genes and 24 were non-classical HLA genes. The researchers noted that the classical HLA gene alleles were highly polymorphic, with, for instance, 20.1 alleles per four-digit-level allele information. The non-classical HLA genes had lower variation levels, 3.1 alleles per four-digit-level allele information.
By visualizing linkage disequilibrium patterns and comparing them, the researchers noted that non-classical HLA genes might have independent genetic landscapes, as compared to classical HLA genes.
Because of that, the researchers developed a new HLA imputation reference panel. They also extended the target regions to include the MHC and its flanking regions, and developed a reference panel using SNP2HLA based on the genotyping of SNPs in the full MHC regions and HLA gene sequence variants. In their cross-validation study, they found this new panel had high accuracy, 96.4 percent and 99.1 percent for four-digit classical and non-classical HLA alleles.
They applied this panel to impute the HLA variants of GWAS genotyped data from 166,190 Japanese individuals collected by the BioBank Japan project.
With that data, the researchers conducted a PheWas of 106 phenotypes, including immune-related diseases, cancers, electrolyte levels, and blood pressure. Half of all the phenotypes they investigated had an association with an MHC allele, they reported. For instance, the researchers uncovered significant non-additive effects of HLA-DPB1*05:01 and HLA-DPB1*02:02 alleles on the risk of Graves' disease.
By clustering the association patterns they saw, the researchers examined the genetic and phenotypic landscape of the MHC regions, which highlighted significant novel MHC associations with 11 traits, such as hyperlipidemia, type 2 diabetes, and creatine kinase, among other phenotypes.
Through fine-mapping, they further resolved some risk variants. For instance, they found that polymorphisms in HLA-DQβ1, HLA-DPα1, and HLA-DQα1 contribute to hepatitis B risk in Japanese individuals.
They also estimated genetic correlations of the MHC region across those phenotypes to uncover shared polygenic architecture. Rheumatoid arthritis, for example, had positive correlations with asthma, diabetes, and bilirubin levels, but had negative correlations with BMI in classical HLA gene variants, while in other MHC variants, it showed negative correlations with hyperlipidemia, stable angina, and more.
This suggested to the researchers that the polygenic architecture of the MHC provides pleiotropic diversity.