the results of Augur

[crystal0721]

Dear, we tested Augur using our data. The code is as follows:

seurat.obj<-readRDS('ZSH2vsZSH1_combined.rds')

augur = calculate_auc(seurat.obj,cell_type_col = "seurat_clusters", label_col = "stim")

seurat_clusters: the cluster type label_col：sample

while the output of augur$AUC as follows： cell_type auc 1 2 0.778726379440665 2 0 0.679890400604686 3 4 0.662003023431595 4 3 0.592526455026455

we found that the cell_type had no cluster 1. we want to know the reason. Thanks for your help.

[skinnider]

I am not sure I understand your question... what exactly is the issue?

[crystal0721]

We had five cluster:0、1、2、3、4； while the results of augur$AUC only had four cluster， there was no auc score of cluster 1.

[skinnider]

How many cells from cluster 1 are in each of the two experimental conditions?

[crystal0721]

The one condition only has 12 cells，because this reason. We want to know how many cells are needed at least?"

[skinnider]

20/condition

On May 17, 2023, at 3:05 AM, crystal0721 @.***> wrote:

 The one condition only has 12 cells，because this reason. We want to know how many cells are needed at least?"

— Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you commented.

[mluciarr]

Hi, thank you for developing this package! Unfortunately, I encountered the same problem. Even though all clusters have more than 20 cells, the last cluster doesn't appear, which is extremely weird, and it does the same with other cases. In my case, I have 6 clusters (0 to 5), but the last cluster (cluster 5) doesn't appear. What could be the reason for that? Thanks a lot in advance.

Lucia R ` augur <- calculate_auc(input = seurat@assays$RNA@counts, meta = data.frame( label = seurat@meta.data$sample, cell_type = seurat@meta.data$cluster))

augur$AUC `

A tibble: 5 × 2 cell_type auc

1 1 0.615 2 2 0.615 3 3 0.600 4 0 0.590 5 4 0.575

[skinnider]

@mluciarr could you share your dataset? @AlanTeoYueYang can you or Laurent investigate?

[skinnider]

One more question - are there 20 cells per condition in the missing cluster?

[mluciarr]

@mluciarr could you share your dataset? @AlanTeoYueYang can you or Laurent investigate?

Hi @skinnider, actually, I've already found the mistake. While discussing this result with one of my colleagues, I noticed that even though cluster 5 (which I referred to as 'cell.type') has 80 cells, there are 14 cells from one of the two labels (T1 and T2). Therefore, the AUC wasn't calculated for this cluster.

I tried it with another resolution, and it worked perfectly. Thank you so much for your rapid response!! Best wishes

[ZhangRuifen]

One more question - are there 20 cells per condition in the missing cluster?

Hi! The Augur helps me a lot! But I have a quesion about the number of cells. I have the same question as below. And I found that it is caused by my cells number which is 10 cells. My question is, could I change the "subsample_size= 5" ? If I changed it whether the result is reasonable? If not reasonable what should I do with its?