An imprinting map of the human placenta based on the application of a novel population genetics approach to RNAseq data. C. T. Watson1, O. Rodriguez1, B. Jadhav1, N. Azam1, D. J. Ho1, K. Cheung1, D. Sachs1, K. Hao1, R. J. Wright2, E. E. Schadt1, A. J. Sharp1 1) Department of Genetics and Genomic Sciences, Icahn School of Medicine at Mount Sinai, New York, NY; 2) Department of Pediatrics, Icahn School of Medicine at Mount Sinai, New York, NY.
We have developed a novel approach for identifying imprinted gene expression through analysis of population genotype frequencies in RNAseq data. Under the null model of bi-allelic expression, the observed genotype frequencies for a transcribed SNP are expected to follow classical Hardy-Weinberg equilibrium (HWE). However, where consistent mono-allelic expression occurs, such as at imprinted loci, this will manifest as a significant departure from HWE in mRNA, with a depletion of heterozygotes. We have applied this approach to analyze an RNAseq dataset from 180 placental samples from the National Childrens Study. We identified 1,213 transcribed SNPs with significantly reduced rates of heterozygosity (HWE p<0.001), corresponding to 56 Refseq genes, three large microRNA clusters, and a further 272 intergenic SNPs located outside of annotated RefSeq transcripts. This list includes 18 previously identified imprinted genes. In order to validate imprinted expression of these loci, we harnessed maternal SNP genotypes available from 31 of the placenta in our study. Utilizing a novel phasing approach we were able to assign parental origin to the expressed allele at ~97% of heterozygous sites. Considering those SNPs for which there were multiple informative placentae (n=598), 96% showed monoallelic expression of the same parental allele across all samples, indicating imprinting as the predominant underlying mechanism. Based on whole genome bisulfite sequencing data, we also observed allele-specific methylation associated with many of these transcripts, and performed validations of parental-specific expression and methylation marks in an independent cohort of placenta/mother pairs. These findings demonstrate the robustness of our HWE approach, which provides the first comprehensive map of imprinting in the human placenta. Intriguingly, many of these placental-specific imprinted genes have functions that are consistent with the parental conflict hypothesis, in which it is theorized that imprinting evolved primarily to regulate transfer of nutrients from mother to offspring: we observed novel paternally-expressed genes that positively regulate angiogenesis, and maternally-expressed genes with inhibitory effects on capillary formation. Furthermore, some of these imprinted genes have recently been shown to have altered expression levels in growth-restricted fetuses, suggesting that they represent strong candidates as modifiers of fetal growth.
You may contact the first author (during and after the meeting) at