scverse
cellphonedb ligrec plot
Hi,
Does sq.pl.ligrec support plots similar to cellphoneDB ?
Because when there are many clusters, the interaction plot generated will be very large and hard to save and to see. In this case, the following summary plots are very useful. Different shape/size/color represents the terms among clusters, and the interacting molecules are not necessary to be plotted. Only the summary are needed.
hi @wangjiawen2013 , you are talking about the heatmap on the right? If yes, unfortunately we don't have such visualization and it is no planned. However, it seems doable to do by aggregating results from the function, as it is outlined in this tutorial https://squidpy.readthedocs.io/en/latest/auto_examples/graph/compute_ligrec.html (you'd have to aggregate the "means" and "pvalues" tables, accordingly to whether you are interested in counting interactions by significance)
it's actually really doable, keeping it open for reminder, maybe in the future
it's actually really doable, keeping it open for reminder, maybe in the future
Agreed, we have the private _heatmap implementation that we could use for this. Will look at this later.
CellPhoneDB has been updated to v3, which incorporate spatial information. Does sq.gr.ligrec adopt this version ?
CellPhoneDB has been updated to v3, which incorporate spatial information. Does sq.gr.ligrec adopt this version ?
Unfortunately, I don't know. I'd ask here directly, as we're using omnipath: https://github.com/saezlab/pypath/issues
CellPhoneDB has been updated to v3, which incorporate spatial information. Does sq.gr.ligrec adopt this version ?
sorry am also not familiar with it, in which way does it incorporate spatial information?
please refer to cellphoneDB v3 paper: https://www.nature.com/articles/s41588-021-00972-2 "Mapping the temporal and spatial dynamics of the human endometrium in vivo and in vitro"
Hi, which kind of data does ligrec need ? raw counts or lognormailized counts ? Besides, sq.pl.ligrec save the figures with transparent background, how to set white background ?
hi @wangjiawen2013 ,
I think the way they incorporate spatial information is by simply subsetting the ligrec output with cell types that are known to co-occurr in visium spots, it's more an analysis step rather than a method step. I'd close this for now.