Differenitial analysis and peak-to-gene correlation anlaysis in Single-cell multiomic analysis identifies regulatory programs in mixed-phenotype acute leukemia

[RegnerM2015]

Hi @rcorces and @jgranja24,

I have methodological question about the analyses performed in "Single-cell multiomic analysis identifies regulatory programs in mixed-phenotype acute leukemia"

The differential analysis between cancer and normal used a "nearest-normal" approach. After projecting the cancer cells from each patient onto the normal reference, you find the k-closest nearest normal cells for each MPAL cell.

From the legend of Fig. 2:

Left, scRNA-seq heat map of upregulated genes (LFC > 0.5 and two-sided t test FDR < 0.01) log2(fold changes) comparing MPAL disease subpopulations to closest non-redundant normal cells

How are the non-redudant normal cells defined here? What value of k is used in the k-NN to identify these background cells and how does this k differ from the k used in creating aggregates for peak-to-gene correlation analysis?

I think that k for the peak-to-gene analysis aggregates would have to be sufficiently smaller than the k used in the differential analysis. Otherwise, you may be creating metacells that contain both cancer and normal cells during the peak-to-gene correlation analysis.

Your thoughts on this and feedback would be greatly appreciated! Thank you!

[rcorces]

I'm not able to answer this and @jgranja24 no longer is involved in active maintenance of ArchR. Since this is a publication-specific question that isnt related to ArchR per-se, I would email the corresponding author. Sorry to not be of more help.

[RegnerM2015]

Sounds good! Thank you for getting back to me.