Modeling allele-specific expression from RNA-seq data
Seminar Room 1, Newton Institute
RNA seq data can address a wide range of functional attributes of the expressed genome, including quantitation of levels of expression, patterns of splicing, and differential allelic expression. In the case of differential allelic expression, the assumption is that each gene that is amenable to this analysis is heterozygous in the examined individuals, and a reasonable null hypothesis is that there is a binomial distribution of counts of transcripts of the two alleles with parameter ½. This talk will explore the violations of this model displayed by RNA-seq data, and will seek to identify the causes. Stochastic variation, differential promoter activities, X-chromosome inactivation, and genomic imprinting are some of the biological factors that can be modeled and inferred from RNA-seq data. Interspecific F1 hybrids provide particularly rich for exploring cis-regulatory divergence. Illustrations from human, mouse, horse, donkey, and Drosophila will be given.