Dataset Information

Analysis of public RNA-sequencing data reveals biological consequences of genetic heterogeneity in cell line populations.

ABSTRACT: Meta-analysis of datasets available in public repositories are used to gather and summarise experiments performed across laboratories, as well as to explore consistency of scientific findings. As data quality and biological equivalency across samples may obscure such analyses and consequently their conclusions, we investigated the comparability of 85 public RNA-seq cell line datasets. Thousands of pairwise comparisons of single nucleotide variants in 139 samples revealed variable genetic heterogeneity of the eight cell line populations analysed as well as variable data quality. The H9 and HCT116 cell lines were found to be remarkably stable across laboratories (with median concordances of 99.2% and 98.5%, respectively), in contrast to the highly variable HeLa cells (89.3%). We show that the genetic heterogeneity encountered greatly affects gene expression between same-cell comparisons, highlighting the importance of interrogating the biological equivalency of samples when comparing experimental datasets. Both the number of differentially expressed genes and the expression levels negatively correlate with the genetic heterogeneity. Finally, we demonstrate how comparing genetically heterogeneous datasets affect gene expression analyses and that high dissimilarity between same-cell datasets alters the expression of more than 300 cancer-related genes, which are often the focus of studies using cell lines.

SUBMITTER: Fasterius E

PROVIDER: S-EPMC6060100 | biostudies-literature | 2018 Jul

REPOSITORIES: biostudies-literature

ACCESS DATA

Similar Datasets

Project description:The current paradigm of osteoblast fate is that the majority undergo apoptosis, while some further differentiate into osteocytes and others flatten and cover bone surfaces as bone lining cells. Osteoblasts have been described to exhibit heterogeneous expression of a variety of osteoblast markers at both transcriptional and protein levels. To explore further this heterogeneity and its biological significance, Venus-positive (Venus+) cells expressing the fluorescent protein Venus under the control of the 2.3-kb Col1a1 promoter were isolated from newborn mouse calvariae and subjected to single-cell RNA sequencing. Functional annotation of the genes expressed in 272 Venus+ single cells indicated that Venus+ cells are osteoblasts that can be categorized into four clusters. Of these, three clusters (clusters 1 to 3) exhibited similarities in their expression of osteoblast markers, while one (cluster 4) was distinctly different. We identified a total of 1920 cluster-specific genes and pseudotime ordering analyses based on established concepts and known markers showed that clusters 1 to 3 captured osteoblasts at different maturational stages. Analysis of gene co-expression networks showed that genes involved in protein synthesis and protein trafficking between endoplasmic reticulum (ER) and Golgi are active in these clusters. However, the cells in these clusters were also defined by extensive heterogeneity of gene expression, independently of maturational stage. Cells of cluster 4 expressed Cd34 and Cxcl12 with relatively lower levels of osteoblast markers, suggesting that this cell type differs from actively bone-forming osteoblasts and retain or reacquire progenitor properties. Based on expression and machine learning analyses of the transcriptomes of individual osteoblasts, we also identified genes that may be useful as new markers of osteoblast maturational stages. Taken together, our data show much more extensive heterogeneity of osteoblasts than previously documented, with gene profiles supporting diversity of osteoblast functional activities and developmental fates. © 2021 The Authors. JBMR Plus published by Wiley Periodicals LLC on behalf of American Society for Bone and Mineral Research.

Project description:Skeletal muscle from meat-producing livestock such as cattle is a major source of food for humans. To improve skeletal muscle growth efficiency or quality in cattle, it is necessary to understand the genetic and physiological mechanisms that govern skeletal muscle composition, development, and growth. Satellite cells are the myogenic progenitor cells in postnatal skeletal muscle. In this study we analyzed the composition of bovine satellite cells with single-cell RNA sequencing (scRNA-seq). We isolated satellite cells from a 2-week-old male calf, cultured them in growth medium for a week, and performed scRNA-seq using the 10x Genomics platform. Deep sequencing of two scRNA-seq libraries constructed from cultured bovine satellite cells yielded 860 million reads. Cell calling analyses revealed that these reads were sequenced from 19,096 individual cells. Clustering analyses indicated that these reads represented 15 cell clusters that differed in gene expression profile. Based on the enriched expression of markers of satellite cells (PAX7 and PAX3), markers of myoblasts (MYOD1, MYF5), and markers of differentiated myoblasts or myocytes (MYOG), three clusters were determined to be satellite cells, two clusters myoblasts, and two clusters myocytes. Gene ontology and trajectory inference analyses indicated that cells in these myogenic clusters differed in proliferation rate and differentiation stage. Two of the remaining clusters were enriched with PDGFRA, a marker of fibro-adipogenic (FAP) cells, the progenitor cells for intramuscular fat, and are therefore considered to be FAP cells. Gene ontology analyses indicated active lipogenesis in one of these two clusters. The identity of the remaining six clusters could not be defined. Overall, the results of this study support the hypothesis that bovine satellite cells are composed of subpopulations that differ in transcriptional and myogenic state. The results of this study also support the hypothesis that intramuscular fat in cattle originates from fibro-adipogenic cells.

Dataset Information

Analysis of public RNA-sequencing data reveals biological consequences of genetic heterogeneity in cell line populations.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets