Differences between Achilles RNAi downloaded files and web portal

Hi, I’m wondering if there is a reason the 2022_Q2 data files I downloaded for RNAi do not match the results from the web portal. For example, the file Achilles_common_essentials.csv includes the gene PRMT5 as a common essential, but PRMT5 is not denoted as common essential on the web portal. I also find that the distribution of gene effect scores in Achilles_gene_effect.csv does not match the distribution shown on the web portal.

I believe there may be some confusion due to which data is derived from RNAi screens and which data is from CRISPR screens.

Achilles_common_essentials.csv and Achilles_gene_effect.csv are data from CRISPR screens which were performed at the Broad. The RNAi data is actually not part of any DepMap release, as we are no longer performing RNAi screens. You can find that data the DEMETER v6 release, under RNAi screens:


So if you look at the distribution of gene effect for these gene, we show the two distributions, purple for RNAi and blue for CRISPR.

The CRISPR screens do report this as common essential, while the RNAi screens show a selective profile. The mechanisms driving knock-out vs knock-down behavior of CRISPR and RNAi are different, and so we tend to think of this not as an artifact, but a reflection of the differences in the mechanism of suppression in the two screens.