Project Summary Alternative splicing generates transcripts that vary between tissues and cell types, and contributes to cell differentiation and organ development. Despite its importance in development and cell identity, alternative splicing is usually neglected in single cell RNA sequencing (scRNA-seq) analysis. This is due to the challenge in identifying confident alternative splicing events from scRNA- seq data, which are limited in read-depth, and comes with large amount of noise due to variation in molecule capture efficiency, amplification bias, and excessive zero-counts or dropouts. We propose a novel approach to infer differential splicing based on short-read full-length scRNA-seq data, utilizing read counts mapping to adjacent exons. In addition, we plan to systematically evaluate existing approaches to answer questions on the strategy regarding the variable, pooling and dispersion estimates. Finally, we will assess the accuracy of short read based methods by comparing against scRNA-seq generated with new types of protocols. The work proposed will increase our understanding on the utility and limitations of short read scRNA-seq data, and provide a better pipeline for alternative splicing analysis in scRNA-seq.