Common Essentials - Agreement between lists

Hello experts!

First of all, I’d like to thank everyone involved in this project for the incredibly useful resource you have provided to the community.

I am interested in carrying out some bespoke analyses of Common Essential genes (e.g. by organelle, by cell lineage etc.) I am however finding it difficult to filter Common Essential genes, as I can’t seem to identify lists of genes that match up. For example, a subset of genes from the Common Essential list (from any of the studies) don’t appear on the Gene Effect list (“CRISPR (DepMap 21Q3 Public+Score, Chronos)”).

I appreciate that much of the numerical data is derived, and the effect scores are dependent on many factors, including post-hoc corrections for various parameters. However, as a general rule to simplify my analysis of whichever gene list I download, would it be sensible to:

  1. Calculate a mean effect score across all cell lines
  2. Use -0.5 as a cut off for common essential/non common essential

Apologies if the above isn’t very clear, and many thanks in advance!

Luke

Hello! Apologies for the confusion, there are multiple common essentials lists in our dataset.

The common_essentials file contains the set of genes used as positive controls, aggregated from two orthogonal knockout studies (Blomen et al. 2015 and Hart et al. 2015). There are a few (~10) genes in this list which do not appear in the gene effect matrices.

Separate from this list, we provide two other common essentials lists: Achilles_common_essentials and CRISPR_common_essentials. These lists are derived from the corresponding Achilles and CRISPR gene effect matrices (ie. without the use of the control essentials list), as described in the Identifying common dependencies section of this preprint: https://www.biorxiv.org/content/10.1101/720243v1.full.

Note that the CRISPR_gene_effect matrix (equivalent to “CRISPR (DepMap 21Q3 Public+Score, Chronos)”) represents data integrated from Achilles and SCORE, meaning it contains only genes which overlap between the two projects. If you were using the common_essentials or Achilles_common_essentials files as your gene lists, that would explain why some genes didn’t match up. Instead, I would suggest using the CRISPR_common_essentials file from the downloads page as your list of common essentials.

Please let me know if I can further clarify anything.

Thank you so much!

Luke