Joint estimation of genetic and parent-of-origin effects using RNA-seq data from human

RNA sequencing allows one to study allelic imbalance of gene expression, which may be due to genetic factors or genomic imprinting. It is desirable to model both genetic and parent-of-origin effects simultaneously to avoid confounding and to improve the power to detect either effect. In a study of experimental cross, separation of genetic and parent-of-origin effects can be achieved by studying reciprocal cross of two inbred strains. In contrast, this task is much more challenging for an outbred population such as human population. To address this challenge, we propose a new framework to combine experimental strategies and novel statistical methods. Specifically, we propose to collect genotype data from family trios as well as RNA-seq data from the children of family trios. We have developed a new statistical method to estimate both genetic and parent-of-origin effects from such data sets. We demonstrated this approach by studying 30 trios of HapMap samples. Our results support some of previous finding of imprinted genes and also recover new candidate imprinted genes.