Possible scATAC extensions

[SarahOuologuem]

Hi,

the following may be possible extensions to panpipes in regards to scATAC-data:

Preprocessing:

• Alternative to using the scanpy sc.pp.highly_variable_genes for the ATAC data: FindTopFeatures() (https://stuartlab.org/signac/reference/findtopfeatures) Signac selects features in this way for the scATAC data, see f.ex.: https://stuartlab.org/signac/articles/pbmc_vignette.html#normalization-and-linear-dimensional-reduction

Visualization:

• Plotting genomic regions: https://stuartlab.org/signac/reference/coverageplot
  They also have many other different plots: https://stuartlab.org/signac/articles/visualization.html
  But I think especially the CoveragePlot() is really nice

Analysis:
Not sure how deep panpipes wants to go on the analysis part, but:

• Signac's gene activity assay
• Find co-accessible networks with Cicero. The co-accessible links can then also be nicely plotted with CoveragePlot(), see: https://stuartlab.org/signac/articles/cicero.html
• Motif analysis

RNA+ATAC:

• For a multiomic dataset it could be interesting to run weighted nearest neighbors and with this graph run UMAP to build a joint visualization, see f.ex.: https://stuartlab.org/signac/articles/pbmc_multiomic.html
  In signac, they use the FindMultiModalNeighbors() function for this