Dimension-reduction methods, such as t-SNE or UMAP, are widely used when exploring high-dimensional data describing many entities, for example, RNA-seq data for many single cells. However, dimension reduction is commonly prone to introducing artifacts, and we hence need means to see where a dimension-reduced embedding is a faithful representation of the local neighborhood and where it is not. We present Sleepwalk, a simple but powerful tool that allows the user to interactively explore an embedding, using color to depict original or any other distances from all points to the cell under the mouse cursor. We show how this approach not only highlights distortions but also reveals otherwise hidden characteristics of the data, and how Sleepwalk's comparative modes help integrate multisample data and understand differences between embedding and preprocessing methods. Sleepwalk is a versatile and intuitive tool that unlocks the full power of dimension reduction and will be of value not only in single-cell RNA-seq but also in any other area with matrix-shaped big data.

中文翻译：

---

使用Sleepwalk探索减少尺寸的嵌入。

在探索描述许多实体的高维数据（例如，许多单个细胞的RNA序列数据）时，广泛使用降维方法（例如t-SNE或UMAP）。但是，降维通常易于引入伪影，因此我们需要一种方法来查看降维嵌入在哪里忠实表示本地邻域，而在哪里则不是。我们介绍了Sleepwalk，这是一个简单但功能强大的工具，它使用户可以交互地探索嵌入，使用颜色描绘从所有点到鼠标光标下的单元格的原始距离或任何其他距离。我们将展示这种方法不仅可以突出失真，还可以揭示数据的其他隐藏特征，以及Sleepwalk如何 的比较模式有助于整合多样本数据，并了解嵌入和预处理方法之间的差异。Sleepwalk是一种多功能且直观的工具，可以充分发挥降维的全部功能，不仅在单细胞RNA序列中而且在具有矩阵状大数据的任何其他区域中都具有价值。