!Significant differences in avg_log2FC between Seurat versions 4 and 5 in FindMarkers function

[xinmiaoyan]

I ran the FindMarkers function on the same data with the same group assigned. However, the avg_log2FC is strikingly different between the old Seurat version 4 and the new version 5. The p-value and adjusted p-value seem the same as before, but the avg_log2FC is significantly different (see attached). Does anyone have a solution to this issue?

[morallawwithin]

Look here:
#6701

[mhkowalski]

Hello,

Also linking our NEWS page that describes this change. The difference is related to choice in pseudocount, which will be much higher for lowly expressed genes.