DepMap 25Q3 vs 25Q2 release: what should I trust?

Q&A
1

I can see a lot of hits in my top 100 co-dependencies in the 25Q2 release now disappeared from the list in the 25Q3 version. These hits had a positive Pearson with p-value indicated significance in 25Q2 but now appear with a negative Pearson in 25Q3. Which dataset should I trust?

2

Would you be willing to share some examples of genes with this behavior? That will be very helpful for us to investigate.

Thanks,

Phil