Nancy Zhang University of Pennsylvania
Traditional gene expression measurements were made on bulk tissue samples containing populations of cells. Recent laboratory advances have made possible the measurement of RNA levels in single cells. This new frontier offers exciting challenges and opportunities. I will describe some of the statistical challenges and propose a new error model for single cell RNAseq that explicitly addresses the technical issues of dropout, amplification artifact, and cell size confounding. I will show that how this model can improve the accuracy of downstream analyses such as differential expression testing. I will also show how to use this model to better characterize the stochasticity of gene transcription at the allele level, and thus improve our understanding of gene regulation. This is joint work with Cheng Jia, Yuchao Jiang, and Mingyao Li.