More information needed about the change of method in Fold Change calculation

[bellenger-l]

Hello !

We compare how you calculate the avg_log2FC between Seurat v4 and v5, but we wanted to have a more detailed explanation why you made this change. We saw this discussion #8229 and this issue #8168, but it's been labelled as answered/closed and it didn't give us an answer.

In v4, the calcul is :

$$avgLog2FC = log2(sum_{group1}/nbrCell_{group1} +1) - log2(sum_{group2}/nbrCell_{group2} +1)$$

In v5, the calcul is :

$$avgLog2FC = log2( (sum_{group1} + 1) / nbrCell_{group1}) - log2( (sum_{group2} + 1) / nbrCell_{group2})$$

Except that we don't understand your explanation in the release note

In addition, in Seurat v5 we implement a pseudocount (when calculating log-FC) at the group level instead of the cell level. As a result, users will observe higher logFC estimates in v5 - but should note that these estimates may be more unstable - particularly for genes that are very lowly expressed in one of the two groups. We gratefully acknowledge feedback from the McCarthy and Pachter labs on this topic.

We saw that in v5 the pseudocount impacts less the original count. Is there a special reason why you made these changes ? Because it's not discussed in your latest paper Hao et al., 2023 and this has a huge effect on the avg_log2FC of poorly expressed genes. Speaking of that, why do you decrease the default value of min.pct parameter if the change in avg_log2FC makes it more unstable for these genes ?

By the way, you mention the feedback from the McCarthy and Pachter labs on this topic., but unfortunately we didn't succeed to find it. Can you please point this discussion ?

We really wanted to understand why these changes were made in order to be fully confident in our analyses and the results' interpretations.

Thanks a lot for your time and package
Best
Lea & @Ning-L