This repository contains the R codes which could reproduce the results of "LTMG: A statistical model of transcriptional regulatory states in single cell RNA-Seq data" by Changlin Wan, Wennan Chang, Yu Zhang, Fenil Shah, Sha Cao, Melissa L. Fishel, Qin Ma, and Chi Zhang.
LTMGSCA is a left truncated mixture Gaussian (LTMG) model-based single cell RNA-seq analysis pipeline that can accurately infer the modality and distribution of individual gene’s expression profile in scRNA-seq data, while decrease the impact from dropout and low expression value. Enabled by LTMG, a differential expression test and a gene co-regulation module identification method, namely LTMG-DGE and LTMG-GCR, are further developed and incorporated within the framework.
We compared LTMG with other state-of-art scRNA-seq models on a comprehensive set of human scRNA-seq data. LTMG in general achieved best goodness of fitting among them (lower KS value means better fitting performance). More detailed illustration could be found at https://www.biorxiv.org/content/early/2018/09/29/430009.
To install the
LTMGSCA demonstration package, you will need to install the release version of
devtools from CRAN with