Meta-analysis of scRNA seq data to define the landscape of human liver immune homeostasis

Single-cell RNA sequencing (scRNA seq) generates highly dimensional transcriptome data to understand cellular phenotypes. Due to high cost and volume of data obtained from these experiments, most studies analyze few unique samples. Pre-analytical tissue processing and different sequencing techniques can alter quantity and phenotype of cells being examined. This study used existing human liver scRNA seq datasets to determine if heterogeneity exists between datasets and if these data could be integrated to define phenotypes across leukocyte subpopulations in healthy human liver. CD45+ cells from three published studies reporting scRNA seq data from human livers were analyzed and leukocyte subpopulations were examined. All three datasets co-clustered, but with differing cell proportions. Expression correlation demonstrated similarity across all studies. Detailed analysis of differential expression identified meta-signatures for each hepatic immune subpopulation. Herein, we introduce a novel and rigorous framework for meta-analysis of scRNA seq datasets and highlight profiles of liver immune homeostasis.