Thanks very much for this amazing tool. So far I have been having really good results with my data, but I would like to move forwards the tutorial you provide.
Once I have managed to align my manifold, I use the latent variables to embed my data on a UMAP, which in my case is more useful than t-SNE. In order to annotate my manifold, I would like to analyse the motifs that are specific for each cluster. Do you have a recommendation on how to go about this?
Next, I would like to use the output from SCALE to integrate it with scRNA-Seq data. Is this something that could be done in SCALE?
Any advice on how to proceed with the analysis will be much appreciated.
UMAP is enabled in the code, replace the method parameters 'TSNE' to 'UMAP' "scale.plot.plot_embedding" function, however, the option is not available to the user in SCALE.py. We will add the option of UMAP soon.
To find the cluster-specific motifs, there are two choices: a. you can use the imputed data and perform chromVAR to find the cell type specific motifs in chromVAR. Embeddings of motifs matrix generated by chromVAR with imputed data is similar to latent embedding of SCALE. b, we provide a function to calculate the cluser-specific peaks (see tutorial), then you can apply motif enrich software like HOMER to find the enriched motifs with these peaks.
You mean the two kinds of omics from the same cell or different cells? If the former, you can concat the scRNA-seq data and scATAC-seq data of the same cell together then apply SCALE. If the latter, SCALE is not capable to tackle two sets of datasets together, though SCALE can handle scRNA-seq individully.
Hi!
Thanks very much for this amazing tool. So far I have been having really good results with my data, but I would like to move forwards the tutorial you provide.
Once I have managed to align my manifold, I use the latent variables to embed my data on a UMAP, which in my case is more useful than t-SNE. In order to annotate my manifold, I would like to analyse the motifs that are specific for each cluster. Do you have a recommendation on how to go about this?
Next, I would like to use the output from
SCALE
to integrate it with scRNA-Seq data. Is this something that could be done inSCALE
?Any advice on how to proceed with the analysis will be much appreciated.
Regards