Entering edit mode
4 months ago
reddyornah
•
0
I am really new to bioinformatics and seurat so I have been analyzing some data for a while. The lab I work in uses Seurat's FindTransferAnchors with a reference data set of healthy mouse tissue and adds the data as a column of predicted cells types. I trusted it until I changed some quality control limits (features) and some new cell types came up with pretty different counts for each. I got concerned so I did some manual annotations based on biomarkers and added module scores to reidentify them. I compared the two predictions and there are a lot of cells that aren't as predicted. The query object i'm using is infected mouse tissue so im wondering if that's why so many cells are being mislabeled.
