igrabski
commented
10 months ago Hi, thanks for your question. There had been an issue with metadata and AggregateExpression in the initial Seurat 5.0.0 release, which was fixed in Seurat 5.0.1. I would recommend upgrading Seurat and seeing if the issue is resolved now. I tried to reproduce this issue with the ifnb data using Seurat 5.0.1, and I get the expected behavior, e.g.:
> pseudo_ifnb <- AggregateExpression(ifnb, assays = "RNA", return.seurat = T, group.by = c("stim","seurat_annotations"))
> head(pseudo_ifnb)
orig.ident stim seurat_annotations
CTRL_CD14 Mono CTRL_CD14 Mono CTRL CD14 Mono
CTRL_CD4 Naive T CTRL_CD4 Naive T CTRL CD4 Naive T
CTRL_CD4 Memory T CTRL_CD4 Memory T CTRL CD4 Memory T
CTRL_CD16 Mono CTRL_CD16 Mono CTRL CD16 Mono
CTRL_B CTRL_B CTRL B
CTRL_CD8 T CTRL_CD8 T CTRL CD8 T
CTRL_T activated CTRL_T activated CTRL T activated
CTRL_NK CTRL_NK CTRL NK
CTRL_DC CTRL_DC CTRL DC
CTRL_B Activated CTRL_B Activated CTRL B ActivatedI will close this issue for now, but please feel free to re-open if upgrading Seurat does not address this problem!
drropadope
shanshancode
I am trying to use AggregateExpression to do DE analysis on pseudobulked data. However, I am having a lot of difficulty getting the output I (think) I should be getting. The meta data objects I use for the group.by argument clearly exist in the object I give (see tables generated in code below), and the AggregateExpression also seems to be finding that meta data because the names of the groups are what I would expect them to be, but for some reason it isn't actually grouping the counts the way it should be- it only creates one column in the meta data matrix instead of the three I think it should be.
In the code below, "cb" is a seurat object. It contains two biological replicates (stored in orig.ident) run as separate libraries. For each biological replicate, there are also 5 conditions that were multiplexed using hashtag labelling. The cb object has been integrated using Harmony and the 5 conditions have been demultiplexed using HTODemux, so the condition labels are stored in HTO_classification in meta data.
I'm sure I'm missing something, but I can't figure out what could be wrong about my input. Any help would be greatly appreciated!
As a quick addendum, I just tried running this command using the Seurat vignette with the ifnb data, and I get the same issue. So it doesn't seem to be specific to my object.