Cytosplore Transcriptomics incorporates state-of-the art clustering and dimensionality reduction techniques for interactive exploration of cell types. One can visualize transcriptome-wide gene expression in combination with metadata of individual cells, perform differential analyses and statistics between manual selections of cells. Using GPU accelarated tSNE and HSNE Cytosplore Transcriptomics is scalable to large datasets having millions of cells, and can interactively produce low-dimensional visualizations of the data hierarchy providing different levels of details.

Cytosplore Transcriptomics is based on Cytosplore and provides many of the analysis tools presented in the previous Cytosplore publications and is available for download. Cytosplore Transcriptomics makes use of HDILib for fast tSNE and HSNE computations.

Cytosplore Transcriptomics is being developed by a team consisting of members from the Leiden Computational Biology Center, the Division of Image Processing at the Leiden University Medical Center, and the Computer Graphics and Visualization Group at the TU Delft

This work was partially supported by an NWO Gravitation grant:BRAINSCAPES: A Roadmap from Neurogenetics to Neurobiology (NWO: 024.004.012) and NWO AES grants 3DOMICS (NWO: 17126), VANPIRE (NWO: 12720) and Genes in Space (NWO: 12721).