PCA/ UMAP bug
Notebook 3: For small data matrices such as CITEseq, the umap wrapper for batch correction fails
sc.pp.pca(adata, svd_solver='arpack', n_comps=50, use_highly_variable=True)
--> ValueError: `k` must be an integer satisfying `0 < k < min(A.shape)`
Can be solved by setting n_comps to a smaller number, BUT ends with:
analyser.wrap_umap(batch_corrections.values(), threads=threads)
--> X_pca does not have enough Dimensions. Provide a Representation with equal or more dimensions than`n_pcs` or lower `n_pcs`
solved by calculation umap by hand
IDEA: Check n_components dynamically to max, and set n_pcs for smaller datasets, such as given in citeseq
Edited by Hendrik Schultheis