High-dimensional single-cell technologies, such as multicolor flow cytometry,  mass cytometry, and image cytometry, can measure dozens of parameters at the single-cell level. FCS Express integrates t-Distributed Stochastic Neighbor Embedding, otherwise known as t-SNE, which is a tool that allows you to map high-dimensional cytometry data onto a two dimension plot while conserving the original high-dimensional structure to help you visualize and analyze high-dimensional data.


     viSNE, which uses the Barnes-Hut implementation of the t-SNE algorithm, was developed by El-ad David Amir and colleagues in 2013 and is based on the t-Distributed Stochastic Neighbor Embedding (t-SNE) algorithm developed by Laurens van der Maaten and Geoffrey Hinton in 2008.


    Briefly, the optimization algorithm included in viSNE searches for the best low-dimensional (e.g. two or three dimensions) representation of the high-dimensional events, so that the pairwise distances between events are best conserved between the high- and the low-dimensional space.


     The final result of the algorithms in FCS Express is a 2D plot in which the positions of cells reflect their proximity in their original high-dimensional space. Plots can further be colored with density or heat mapping of each parameter allowing for easy visualization of populations. Gates may also be used to define populations of interest and quickly explored across each parameter using parameter overlays on histograms.


     Given that t-SNE is an highly demanding algorithm, in 2014 Laurens van der Maaten developed a faster version of the tSNE algorithm exploiting the Barnes-Hut algorithm. The Barnes-Hut method gives an approximate result of tSNE but it is faster than the exact tSNE. FCS Express allows the user to use both the exact tSNE algorithm and the Barnes-Hut approximation. Moreover, the developers team at De Novo Software worked hard to make both the exact algorithm and the Barnes-Hut approximation even faster. You can find a performance test in the next chapter.


    In 2019, Anna Belkina and colleagues further improve the performance of t-SNE by adjusting a series of hard-coded settings of the original algorithm resulting in a new implementation named Opt-SNE which is available in FCS Express.  


    For a more comprehensive understanding of t-SNE, please click here.



More details on viSNE/tSNE can be found in the following publications:

Van der Maaten, L. Accelerating t-SNE using Tree-Based Algorithms. J. Mach. Learn. Res. 15, 3221-3245 (2014).

Amir el-AD et al. viSNE enables visualization of high dimensional single-cell data and reveals phenotypic heterogeneity of leukemia. Nat. Biotechnol. 31, 545–552 (2013).

Van der Maaten, L. & Hinton, G. Visualizing data using t-SNE. J. Mach. Learn. Res. 9, 2579–2605 (2008).

Anna C. Belkina et al., Automated optimized parameters for T-distributed stochastic neighbor embedding improve visualization and analysis of large datasets. Nature Communications. 10 (1), 5415 (2019)


Additional resources:

Github by Laurens van der Maaten, co-author of original implementation: https://lvdmaaten.github.io/tsne/

Distill article discussing the effective use of t-SNE: http://distill.pub/2016/misread-tsne/



To begin working with viSNE in FCS Express please see the topics on:


t-SNE performance in FCS Express V6

Defining a tSNE transformation

Applying a tSNe transformation

Working with tSNE transformations

Saving a tSNE transformation