Dear developers,

Thanks for the very useful implementation!

I plan to use your tool for single-cell analysis. In this context, I have different representations of the data. For example, the adjacency matrix can be built as a symmetric nxn KNN graph between cells based on their similarity on the expression of thousands of genes.

However, I'm a bit confused about the choice of the initial position matrix. Therefore, I would appreciate your input about how to select it. For example, whether to use PCA, UMAP, or even the original table on which the adjacency matrix was built on.

Best,
Mohamed