The art of using t-SNE for single-cell transcriptomics.
Kobak D, Berens P. · 2019
Read The art of using t-SNE for single-cell transcriptomics. by Kobak D, Berens P. (2019) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/europmc-pmc6882829.txt.
Description
Kobak and Berens offer a practical guide to a popular plotting method, in Nature Communications. Single-cell studies measure thousands of genes in up to millions of cells. To see the data, analysts squash it into two dimensions with t-SNE. The method reveals local groupings beautifully. Used naively, it also distorts the big picture, inventing gaps and hiding real distances.
The authors show how to avoid the worst traps. Starting the layout from a principal component map, using a high learning rate, and mixing similarity scales all help. For very large datasets they add exaggeration and a downsampled start. Worked examples on published data show the same cells forming far more faithful maps.
Catalog record
| CREATOR | Kobak D, Berens P. |
|---|---|
| YEAR | 2019 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 77 KB |
| LICENSE | CC BY 4.0 |
| RIGHTS | CC BY |
| SOURCE | Europe PMC |