Efficient unbiased data analysis is a major challenge for laboratories handling large flow and mass cytometry datasets. We present EmbedSOM, a non-linear embedding algorithm based on FlowSOM that improves the analysis by providing high-performance embedding method for the cytometry data. The algorithm is designed for linear scaling with number of data points, and speed suitable for interactive analysis of millions of cells without downsampling. At the same time, the visualization quality of single cell distribution within cellular populations and their transition states is competitive with the current state-of-the-art algorithms. We demonstrate EmbedSOM properties on two use-cases, showing benefits of using the interactive algorithm speed in supervised hierarchical dissection of cell populations, and the scalability improvement by efficiently processing very large datasets.
Kratochvíl, M., Koladiya, A., Balounova, J., Novosadova, V., Sedlacek, R., Fišer, K., … Drbal, K. (2018, December 20). SOM-based embedding improves efficiency of high-dimensional cytometry data analysis. BioRxiv. bioRxiv. https://doi.org/10.1101/496869