Amir, E.-a.D. et al. Nat. Biotechnol. 31, 545–552 (2013).

As technologies mature that can monitor many parameters simultaneously on single cells, visualization of the resulting data is becoming a bottleneck. One may be able to label heterogeneous cellular populations for tens of surface markers using mass cytometry, for instance, but methods to interpret the signals in terms of the underlying population structure are still limited. Amir et al. now report viSNE, a tool for the interpretation and visualization of high-dimensional data, in which single-cell resolution is maintained. The algorithm projects the data from high-dimensional to two-dimensional (2D) space; in the resulting 2D scatter plot, each point represents the position and relationships of a cell in high-dimensional space. Applied to normal human bone marrow, viSNE reproducibly returns expected cell subpopulations; leukemic bone marrow, by contrast, shows a comparatively abnormal structure.