Single cell-based computational tool for subpopulation conversion

Responsive Image


Convert between cell subtypes by using single-cell data to identify an optimal, subtype-specific set of core transcription factors personalized for your own human or mouse cell study.

File formats
  • Starting cell population: Select or upload a tab-separated and unquoted file with the single-cell gene expression data from the cell type you started your differentiation protocol from. Genes should be labeled according to the Gene Symbols nomenclature.
    For optimal results, the platform used to sequence the starting and the target cell types should be the same.
  • Gene expression matrix: Upload a tab-separated and unquoted file of your raw (unnormalized) single-cell data where each column represents a cell with an unique name and the rows contain gene expression values. Genes should be labeled according to the Gene Symbols nomenclature.
    For a faster upload, please consider submitting a gene expression matrix containing only transcription factors or run the code locally [repository here].
  • Cell annotation: Upload a tab-separated and unquoted file where the first row matches the unique cell names in the gene expression matrix and the second row corresponds to the cluster to which each cell belongs. Please make sure to remove any whitespaces and special characters from your cluster names. For analysing two or more clusters as one, check the box “I would like to merge the selected subpopulations”.

Examples: Please select "Preview Examples" button in the main page to find gene expression matrix and cluster information format examples, obtained from La Manno et al. 2016 data.


The results are compressed in a .zip file. Inside, you will find a summary file that contains the metadata of your analysis, an hierarchical clustering dendrogram of your gene expression matrix, and two tables: "cores.tsv” containing the predicted cell conversion TFs for each target subpopulation, ranked by the fold-change; and “markers.csv” containing the top 10 predicted marker genes of each target subpopulation, ranked by JSD score (low JSD = unique and specific marker).


This software was developed in the Computational Biology Group by Mariana Ribeiro, Dr. Satoshi Owaka and Prof. Dr. Antonio del Sol.


Within any publication that uses any methods or results derived from or inspired by TransSynW, please cite:
Ribeiro MM, Okawa S, del Sol A. TransSynW: A single-cell RNA-sequencing based web application to guide cell conversion experiments. STEM CELLS Transl Med. 2020;1–9.


By clicking "Submit", you agree to our Terms of Use.