UMAP 將會與tsne一樣作為高緯數(shù)據(jù)可視化的利器，并且優(yōu)于tsne.
- UMAP 與 t-SNE 均是非線性降維，多用于數(shù)據(jù)可視化
- UMAP 結(jié)構(gòu)與t-SNE一致
- UMAP 計算更快
- UMAP 能更好地反映高緯結(jié)構(gòu)，比t-SNE有著更好的連續(xù)性
- 這種連續(xù)性反映到單細(xì)胞分析中就是能更好滴可視化細(xì)胞的分化狀態(tài)（UMAP better represents the multi-branched continuous trajectory of hematopoietic development）
- We conclude that both techniques are similar in their visualisation capabilities, but UMAP has a clear advantage over t-SNE in runtime, making it highly plausible to employ UMAP as an alternative to t-SNE in mIF data analysis.
- Most interestingly, UMAP’s branching was able to highlight biological lineages, especially in identifying potential hybrid tumour cells (HTC)
Evaluation of UMAP as an alternative to t-SNE for single-cell data
Herein we comment on the usefulness of UMAP high-dimensional cytometry and single-cell RNA sequencing, notably highlighting faster runtime and consistency, meaningful organization of cell clusters and preservation of continuums in UMAP compared to t-SNE.
Dimensionality reduction for visualizing single-cell data using UMAP
- It also provides the useful and intuitively pleasing feature that it preserves more of the global structure and, notably, the continuity of the cell subsets.
- In addition to making plots easier to interpret, we note that this also improves its utility for gen-erating hypotheses related to cellular development.
- UMAP outputs are faster to compute compared with Barnes–Hut t-SNE, much faster than scvis, and comparable to FIt-SNE. UMAP embeddings are more reproducible than other methods, notably more so than those from t-SNE implementations.
