A genome-wide approach to identifying novel-imprinted genes

Genomic imprinting is an epigenetic process in which the copy of a gene inherited from one parent (maternal or paternal) is consistently silenced or expressed at a significantly lower level than the copy from the other parent. In an effort to begin a systematic genome-wide screen for imprinted genes, we assayed differential allelic expression (DAE) at 3,877 bi-allelic protein-coding sites located in 2,625 human genes in 67 unrelated individuals using genotyping microarrays. We used the presence of both over- and under-expression of the reference allele compared to the alternate allele to identify candidate-imprinted genes. We found 61 genes with at least twofold DAE plus "flipping" of the more highly expressed allele between reference and alternate across heterozygous samples. Sixteen flipping genes were genotyped and assayed for DAE in an independent data set of lymphoblastoid cell lines from two CEPH pedigrees. We confirmed that PEG10 is paternally expressed, identified one gene (ZNF331) with multiple lines of data indicating it is imprinted, and predicted several additional imprinting candidate genes. Our findings suggest that there are at most several hundred genes in the human genome that are universally imprinted. With samples of mRNA from appropriate tissues and a collection of informative cSNPs, a genome-wide search using this methodology could expand the list of genes that undergo genomic imprinting in a tissue- or temporal-specific manner.