Eric Bonnet

Hi @yuhanH Thanks for your message. I checked the clustering with `FindNeighbors` and it runs fine for `pca` but with `apca` I have the same error message. Code used: ```...

Hi @yuhanH Thanks for your suggestions. It seems that there are no NAs or nan values in the cell embeddings: ``` > Embeddings(CD3, reduction="apca") %>% is.na() %>% sum [1] 0...

So it means that the nearest-neighbor procedure implemented in RunUMAP (via uwot package, if I'm not mistaken) is running fine with this dataset, right ?

Turns out after investigation of the dataset by @yuhanH that the error message was caused by lots of "empty" cells in the ADT (proteins) assay. After removing the empty cells...

Thanks @yuhanH for your help on this case. I now close this bug report.

Thanks ! I'll do that soon.

Done, but same error occurred : ``` $ hicPCA -m mat_160K_corrected.cool --outputFileName pca1.bw pca2.bw pca3.bw --format bigwig Traceback (most recent call last): File "/env/cng/proj/LBA/scratch/eb/tools/anaconda3/envs/explo/bin/hicPCA", line 7, in main() File "/env/cng/proj/LBA/scratch/eb/tools/anaconda3/envs/explo/lib/python3.10/site-packages/hicexplorer/hicPCA.py",...

Thanks Eoghan for the answer. I just have another question. In the preprint there is a graph analysis procedure to select the best alignments. Is the code to do that...

Indeed, with the patch the pipeline runs successfully. Thanks Nicolas for your fast response and for the tip.