Dataset Information

Leveraging the Cell Ontology to classify unseen cell types.

ABSTRACT: Single cell technologies are rapidly generating large amounts of data that enables us to understand biological systems at single-cell resolution. However, joint analysis of datasets generated by independent labs remains challenging due to a lack of consistent terminology to describe cell types. Here, we present OnClass, an algorithm and accompanying software for automatically classifying cells into cell types that are part of the controlled vocabulary that forms the Cell Ontology. A key advantage of OnClass is its capability to classify cells into cell types not present in the training data because it uses the Cell Ontology graph to infer cell type relationships. Furthermore, OnClass can be used to identify marker genes for all the cell ontology categories, regardless of whether the cell types are present or absent in the training data, suggesting that OnClass goes beyond a simple annotation tool for single cell datasets, being the first algorithm capable to identify marker genes specific to all terms of the Cell Ontology and offering the possibility of refining the Cell Ontology using a data-centric approach.

SUBMITTER: Wang S

PROVIDER: S-EPMC8455606 | biostudies-literature |

REPOSITORIES: biostudies-literature

ACCESS DATA

Similar Datasets

Project description:Purpose: To identify distinct gene expression and functional profiles for the three main cell types (epithelial, keratocyte and endothelial) of the human cornea. Methods: RNA-sequencing was performed using total RNA isolated from ex vivo corneal epithelial cells (evCEpC), keratocytes (evK) and endothelial cells (evCEnC) obtained from 3 donor corneas obtained from a commercial eye bank. Transcriptomic analysis was performed using Kallisto (alignment (hg38) and quantification) and Sleuth (differential gene expression(DGE)), with transcript abundances calculated as transcripts per kilobase million (TPM). Expression was defined as TPM≥7.5 and significant DGE as a fold-change ≥4 and a false-discovery rate adjusted p-value≤0.05. Cell type specificity was defined as genes adhering to the above expression and DGE criteria and not expressed (i.e., TPM<7.5) in the other two cell types. Gene ontology enrichment analysis was performed on the cell type-specific gene lists using the gene group functional profiling (g:GOSt) tool within the web-based g:Profiler suite. Results: We identified 205 genes specific in evCEpC, and enrichment of epithelial-associated GO terms (e.g., cornified envelope, epidermis development, cell-cell adhesion). We identified 76 genes specific in evK, and fibroblast-associated GO terms (e.g., collagen metabolic process, metallopeptidase activity, extracellular matrix organization). We identified 96 genes specific in evCEnC, and at least one CEnC-associated GO term (e.g., cellular cation homeostasis) but several synapse-associated GO terms (e.g., synapse organization, synapse part). Conclusions: The human cornea is comprised of three main cell types that play important roles in maintaining corneal clarity and vision. Our results demonstrate that a small subset (0.4-1%) of the protein-coding genes confers distinct and classic functional properties associated with each cell type. In addition, we identified a novel association and significant functional overlap between CEnC and synapses. This may lead to insights into the molecular mechanisms of transendothelial transport and secretion, which are integral metabolic features of the corneal endothelium.

Dataset Information

Leveraging the Cell Ontology to classify unseen cell types.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure