SCRIP: An Accurate Simulator for Single-Cell RNA Sequencing Data

• Activity Number: 55 - Statistical methods for data from single cell technologies
• Type: Contributed
• Date/Time: Sunday, August 8, 2021 : 3:30 PM to 5:20 PM
• Sponsor: Section on Statistics in Genomics and Genetics
• Author(s): Fei Qin* and Xizhi Luo and Feifei Xiao and Guoshuai Cai
• Companies: University of South Carolina and University of South Carolina and University of South Carolina and University of South Carolina
• Keywords: scRNA-seq; simulation

Abstract

With the recent advancements in single-cell RNA sequencing (scRNA-seq), solid simulation methods for scRNA-seq data are required to develop reliable analysis methods. However, due to the noisy nature of scRNA-seq data, current available simulation methods cannot sufficiently capture and simulate important properties of real data, especially the biological variation. In this study, we developed SCRIP, a novel simulator for scRNA-seq data. SCRIP adopted a zero inflated Gamma-Poisson hierarchical model and enabled the simulation of various inherent characteristics of scRNA-seq data including biological dispersion, transcriptional bursting and drop-out event. SCRIP improved scRNA-seq data simulation by incorporating mean-variance dependency via generalized additive model. Also, SCRIP enabled the simulation of bursting effect based on Beta-Poisson mixture distribution. Compared to existing simulators, SCRIP showed a significantly higher accuracy of stimulating key data features including mean-variance dependency in all experiments. SCRIP also outperformed 5 existing methods in capturing between-cell type variation and recovering cell trajectories.