FAQ
CellKb Immune is free for academic users to search cell types matching one or more genes. Commercial users can also signup for a free trial license to evaluate CellKb for a limited period.
Academic and commercial users need to purchase an annual subscription to CellKb to access and download all data. We also provide paid service options to customize CellKb and integrate it with other databases.
Contact us to request a demo and pricing.

CellKb is a knowledgebase of cell type specific signatures of gene expression patterns. It is currently organized into two resources:
1. Cell type signatures from Single-cell RNAseq Datasets
2. Cell type signatures from Bulk RNAseq/microarray Datasets
These resources can be searched using a ranked list of genes or a descriptive keyword.

A dataset in CellKb is a publication with a PubMed ID that describes a single-cell or bulk experiment. It contains one or more cell type signatures. For example, Han et al. (PubMed: 29474909) is a dataset describing the Mouse Cell Atlas identified using Microwell-Seq with 98 cell type signatures in CellKb.

A marker gene set or signature is a cell type specific list of marker genes defined within a dataset. Each signature is assigned to a cell type.
Marker gene sets for cell types (single-cell clusters) are taken as defined by the authors in supplementary tables, where available. They are calculated for each cell type using the Wilcoxon rank sum test, where raw (after normalization) or normalized expression values are available with cell type annotations.

A cell type is a name given to a signature in a dataset based on its biological characteristics as described by the authors of the experiment. Signatures with similar properties are assigned to the same cell type. For example, the cell type "goblet cell" is associated with multiple signatures from one or more datasets.
Cell types in CellKb are given appropriate Cell Ontology and/or Uberon Ontology terms where available. This helps to formalize the cell type nomenclature.

Marker gene sets for 9 species are currently available in CellKb. These are H. sapiens, M. musculus, D. rerio, C. elegans, D. melanogaster, R. norvegicus, M. fascicularis, A. thaliana and M. mulatta. We will continue to add more species as more marker gene sets are published.

Cell and/or Uberon Ontology terms are assigned to cell types, tissues, organs and body structures by manually looking up the most appropriate term(s) based on the biological characteristics described in the publication. Some cell types may not be assigned a Cell/Uberon Ontology term if an appropriate term is not found.

A list of ranked genes can be used to identify matching cell type signatures from single and bulk datasets, or consensus signtures in CellKb using rank-based method. This method takes into consideration the number of genes in the user query and cell type signatures, the number of overlapping genes between the two along with their presence towards the top, middle or bottom of either list.

Signature reliability is calculated as the ratio of the average Spearman's correlation coefficient of the signature with all other signatures of the same cell type and that with all other cell types. A value greater than 1 indicates a reliable signature since it is better correlated with other signatures of the same cell type compared to those of other cell types.

The protein-protein interaction network for query genes is calculated based on experimentally identified interactions in the HitPredict database. Gene identifiers given by the user are mapped to protein UniProt IDs and interactions for these proteins are displayed along with reliability scores calculated by HitPredict. Marker genes of the matching cell type are highlighted in the network.

We are committed to updating CellKb on continuously to include the latest publications identifying cell types using single-cell technologies, add new features, and improve overall user experience and accessibility.

CellKb is developed by Dr. Ashwini Patil with technical support by Ajay Patil. CellKb is licensed through Combinatics Inc., a Tokyo-based bioinformatics company. Prior to founding Combinatics, Ashwini was a Lecturer at the University of Tokyo.