Making colors consistent for the Hertzano/Ament single cell datatsets

[beamilon]

@songeric1107

The colors should be consistent for the cell types between P2 and P7 for the following datasets:

• P2, mouse, scRNA-seq, cochlea (Hertzano/Ament) and P7, Mouse, cochlea, scRNAseq (Hertzano/Ament)

Same for:

• P2, mouse, scRNA-seq, utricle (Hertzano/Ament) and P7, mouse, scRNA-seq, utricle (Hertzano/Ament)

[songeric1107]

@beamilon there are inconsistent cell types between those datasets, what pattern should I follow? (in Ear (diverse variety) profile).

Meanwhile, all the datasets are colored based on users's cell_type_color pattern. I could modify it if Kevin can provide the final matched color pattern of those datasets.

[Kvnrose]

Hello Yang

My Seurat objects that were used did not have specific colors for each cell cluster - I just used the defaults so therefore they will not match between the two datasets.

While were at it, lets also update some of the cell type names: 1) Rgs5+ Cells = Perictyes 2) Pf4+ Cells = Immune Cells 3) Oc90+ Cells = Roof Cells 4) Crabp1+ Cells = GER (Greater Epithelial Ridge) 5) Fst+ Cells = Sulcus Cells 6) Apoa4+ Cells = Sulcus Cells

With these changes there are only a couple of differences in the cell types between P2 and P7:

1) P7 has a extra mesenchyme population - Make Mes_Cells 2/3 (P2) the same color as Mes_Cells 2 (P7) 2) Pillar_Cells from P2 can be the same color as DC_and_PCs at P7 3) Crabp1+ (now GER) is not found at P7 - make it a unique color 4) Dividing_Glial_Cells and Dividing_Mes_Cells are not found at P7 make them a unique color.

All other cell types can have matching colors between the two datasets.

[songeric1107]

@Kvnrose do you want me to replace the celltype name in the original file? or should I upload another version with modified names?

[Kvnrose]

Does not matter, whatever is easier for you!

[beamilon]

@Kvnrose @songeric1107 @jorvis I noticed that the new colors are showing in the curator account but not in my account. Is it a question of ownership for the datasets.

[songeric1107]

@beamilon , could you be more specific? I see the same color for curator / my own account. ![Screen Shot 2022-06-02 at 3 35 48 PM](https://user-images.github

[beamilon]

This is what I see in my account. These are the old colors

[songeric1107]

this is what I see from my account.

[beamilon]

I think the problem comes from the fact that in the "owner displays" the above displays exist only with P7 cochlea and P2 utricle Below is an example of the displays for P7 utricle where it is not an option so I had to create my own display. If you switch displays to the violin plot for example, then you are stuck with creating your own display because the owner doesn't have an equivalent one. I recreated my own display and it still comes with the old colors. I am not sure how it makes sense.

[beamilon]

I just figured it out. The colors for the cell-types and the colors for the louvain don't correspond. So if you choose louvain for "colorize legend by" you get the old colors. The top image is coloring by louvain and the bottom one is coloring by cell-type. So the louvain colors need to be match to the cell type colors if possible