CellKb release notes


CellKb Version 2.01
Release Date: 12 October 2020
Features
  • Content: Version 2.01 of CellKb includes 22,528 marker gene sets representing 1,802 unique cell types from 11 species collected from 403 publications. Cell types from Macaca fascicularis were added in this version.
  • NEW Searching multiple gene clusters in a single query: Users can give multiple gene lists as input to CellKb search to find matching cell types in one query. Users need to specify a cluster or cell type identifier for each gene list that is ranked by significance of the genes.
  • NEW Submitting genes in any column: Users can submit the input with the genes listed in any column. CellKb no longer requires users to submit genes in the first column. Instead, it prompts the user to select the column type before submitting a search query.

CellKb Version 2.0
Release Date: 26 August 2020
Features
  • Content: Version 2.0 of CellKb includes 22,528 marker gene sets representing 1,802 unique cell types from 11 species collected from 403 publications. Cell types from Macaca fascicularis were added in this version.
  • Free Trial license: A trial version has been introduced, that gives access to the browse and search functionality. Academic users can use this version for an extended period. Search results are limited to the top 20 matching cell types. The facility to download results is not available in this license.
  • Browsing cell types: All cell types and their marker gene sets can be viewed with the browsing facility in CellKb. Users can access cell types based on categories such a species, publications, disease, organ/tissue and cell type.
  • Viewing complete marker gene sets: Users can now navigate to a cell type and see its complete marker gene set along with associated values as in the original manuscript.
  • Searching matching cell types: Users can enter a gene list, with associated values such as log fold changes and p-values, to find matching cell type marker gene sets in CellKb. Match score is calculated using a variation of the Rank Bias Overlap method.
  • Searching cell types using a keyword: Cell types can be searched for a particular species using its name, tissue, disease condition, publication details, a gene from its marker set and other experimental conditions. Keyword search suggests context-sensitive keywords to users.
  • Searching cell types across species: Users can give a gene list from one species and select another species to search cell types in. During search for matching cell type, genes from the query cell type are converted to orthologous genes from the target species.
  • Selection of experiment type: Cell type search using gene lists and keywords can be limited to cell types identified in single cell and/or bulk experiments.
  • Download search results: All search results can be downloaded in a tab-delimited format.

CellKb Version 1.0
Release Date: 26 April 2020
Features
  • Content: Version 1.0 of CellKb includes 18,985 marker gene sets representing 1,072 unique cell types from 10 species collected from 316 publications. Most of the data is collected from cell types identified in single-cell experiments. Signatures from some bulk experiments in public databases like MSigDB, Human Protein Atlas and SingleR are included in the database.
  • Searching matching cell types: Users can enter a gene list, with associated values such as log fold changes and p-values, to find matching cell type marker gene sets in CellKb. Match score is calculated using a variation of the Rank Bias Overlap method. Users can search for matching cell types from single-cell experiments and bulk experiments through separate interfaces.
  • Searching cell types using a keyword: Cell types can be searched for a particular species using its name, tissue, disease condition, publication details, a gene from its marker set and other experimental conditions. Keyword search suggests context-sensitive keywords to users.
  • Searching cell types across species: Users can give a gene list from one species and select another species to search cell types in. During search for matching cell type, genes from the query cell type are converted to orthologous genes from the target species.
  • Searching consensus signatures: Consensus signatures were calculated for all unique cell types by combining the marker gene sets from multiple experiments using logistic regression. Users can search the consensus signatures using gene lists or keywords.
  • Download search results: All search results can be downloaded in a tab-delimited format.