Most of the time when people generate tSNE plots of flow data, they're interested in comparing how a treatment or condition affects the phenotype of the cells, and using the tSNE to show this. Putting the two side-by-side allows us to spot differences. How do we know if the distribution of cells on two tSNE plots is actually different, though? Well, here's a method that Carlos Roca developed called the Cross Entropy test.
Learn how you can compare tSNE plots with a robust statistical method!
Code available for academic use here.
Comments