t-SNE is a popular dimensionality reduction method for, among many other things, identifying transcriptional subpopulations from single-cell RNA-seq data. However, the sensitivities of results to and the appropriateness of different parameters used have not been thoroughly investigated. Building on previously developed Javascript-based implementations of the t-SNE algorithm, we present this user-friendly web tool to begin to interactively explore the impact of these parameters on identified subpopulations.

Try it out ☞ jefworks.github.io/tsne-online/