The landscape of genomic imprinting across diverse adult human tissues

Abstract

Genomic imprinting is an important regulatory mechanism that silences one of the parental copies of a gene. To systematically characterize this phenomenon, we analyze tissue specificity of imprinting from allelic expression data in 1582 primary tissue samples from 178 individuals from the Genotype-Tissue Expression (GTEx) project. We characterize imprinting in 42 genes, including both novel and previously identified genes. Tissue specificity of imprinting is widespread, and gender-specific effects are revealed in a small number of genes in muscle with stronger imprinting in males. IGF2 shows maternal expression in the brain instead of the canonical paternal expression elsewhere. Imprinting appears to have only a subtle impact on tissue-specific expression levels, with genes lacking a systematic expression difference between tissues with imprinted and biallelic expression. In summary, our systematic characterization of imprinting in adult tissues highlights variation in imprinting between genes, individuals, and tissues.

Figure 1.

Examples of allelic expression patters. In panels A and B, each dot represents RNA-seq haplotype counts of an individual, summed up over phased heterozygous sites across the gene, and in C–E, each dot is a SNP in an individual. (A) The strong monoallelic expression of DLK1 supports its previously known status as an imprinted gene. (B) MEST is almost fully imprinted in lung but biallelic in testis. (C) In PAX8, some individuals show (nearly) monoallelic expression, while others are biallelic. It is an example of a gene that has been excluded from our list of imprinted genes due to the high heterogeneity of allelic expression, which could be due to variable imprinting, cis-regulatory variants, or other effects. (D) MAP2K3 in Geuvadis LCLs has substantial monoallelic expression without additional genotype quality filters on 1000 Genomes data. In fact, many of the SNPs are likely to be truly homozygous, since they fail the HWE test, and are removed from the final analysis. (E) UQCRFS1 in Geuvadis LCLs shows a pattern where all monoallelic sites have only the alternative allele present. This pattern is not consistent with imprinting, and the gene will be filtered out by the “flip test” requiring observation of monoallelic expression of both alleles.

Figure 2.

Imprinting across tissues for the 42 genes detected as imprinted. The color denotes τ, the average ratio of the higher expressed allele to the total read count. See Table 1 for tissue abbreviations. (**) Previously confidently imprinted genes, (*) provisionally identified imprinted genes.

Figure 3.

Variation in imprinting. (A) The number of tissues in which genes are imprinted or biallelic for maternally and paternally expressed genes. (B) Sex-specific imprinting in muscle, where females have lower median τ than males, measured across all genes identified as imprinted in muscle. Each data point corresponds to an individual. (C) An example of variation of imprinting between individuals in ZNF331, with color denoting τ (see Fig. 2). (D) Median expression level of genes in tissues where they are imprinted versus biallelic (see also Supplemental Fig. S20). Only genes with both imprinted and biallelic tissues are shown.

Figure 4.

Tissue differences in the expressed allele. The figure shows comparison of the reference allele ratios of the same SNPs in the same individuals in brain and muscle. ZDBF2 is an example of the typical pattern of the same expressed allele in the two tissues; in IGF2, brain expresses a different allele than muscle, and GRB10 is strongly imprinted only in brain but has a slight signal of muscle expression from the opposite allele than in the brain. All the correlations are significant (p < 0.005).