emperor
emperor copied to clipboard
Suggestions: layered selections
Hi EMPeror lord,
For biplots, it would be great to be able to subset visualization items based on multiple criteria, for example: first remove features' arrows for e.g. "taxon level 1" and then e.g. ">10" in sample prevalence.
This would help one willing to bypass the limitation of being restricted to the top 5-rank arrows and have the possibility to subset using a combination of criteria.
--> You did propose a "categories merging" feature that would allow selecting based on two or more metadata merged "on the fly". Yet I think this could be tedious since the number of factors in the merged column could explose rapidly.
The way I do this currently is to show all the arrows using the "--p-number-of-features" options set to the number of features and then subset using a metadata column tailored for my needs - or editing the ordination.txt to only keep the arrows I want.
One way I see around woulda be to allow a default start with no arrow and then removing using a first metadata factor, then being able to add another dropdown menu, then removing using a second metadata factor from this additional menu, etc... Just like for the wonderful "HivePanelExplorer": https://github.com/hallamlab/HivePanelExplorer/wiki/Hive-Panel-Explorer-Beginner's-Guide#exploring-your-hive-panel-1
Also, could a "Visibility" selection be based on numeric threshold? (I'm gonna raise another issue as this is seperate I guess).
Thanks! F