Finding and exploring reproducible cell phenotypes with the Stemformatics data portal

Abstract Stemformatics is an established online data portal which hosts hundreds of curated gene expression datasets. It has been serving the stem cell research community for over a decade, by hosting transcriptional profiles of pluripotent and adult stem cells and their progeny from multiple tissues and derivation methods. The portal provides easy-to-use online tools to explore gene expression patterns in published data. In recent years, Stemformatics has shifted its focus from curation to collation and integration of public data with shared phenotypes. It now hosts several integrated expression atlases based on human myeloid cells, which allow for easy cross-dataset comparisons and discovery of emerging cell subsets and activation properties. The inclusion of laboratory-derived cell types enables users to benchmark their own data, to assist with cell-type standardisation or improve cell-derivation methods. The sample annotations have been greatly improved to enable better data integration, and the website has also undergone a major upgrade to modernise its visualisation tools. An application programming interface server also provides the data directly for computational users. Stemformatics is an open-source project and readily available at stemformatics.org.