Hi there, sorry if this question is naive.
when looking at the co-dependency list for an essential gene. What are the chances that the (top) hits are in the list simply because they are also essential genes? Or is there a way the data is corrected for this and it selects for functional co-dependency?
thanks in advance!
This is a great question. In general, there is a risk of confounding factors (like overall screen quality) having a significant impact on measured correlations among common-essential genes in particular (See for example our paper here for descriptions of this with RNAi data, and similar issues can arise with CRISPR data).
We believe that our latest data processing methods do a good job of removing these biases in order to maximize the ability to identify meaningful relationships among common-essential genes, but it’s hard to rule out such effects.
In general, we and others do find that there is quite a bit of biologically meaningful signal present among co-dependency relationships for common-essential genes. So I’d probably recommend using this info, but keep potential limitations of the data in mind particularly when assessing patterns within the common-essential gene sets as the results can be more dependent on the choice of data normalization and processing methods.
Thank you very much, this response is very helpful!
