Identification of tumor suppressor genes

Q&A
1

Hi,

I have a question about gene effect score and gene dependency score.

  1. Does essentiality only mean that knocking out a gene will cause cancer cells to stop growing or even die? Is there a reverse concept, that knocking out a gene can cause cancer cells to proliferate?
    We know that the more negative the gene effect score, the more likely the gene is to promote tumor growth. So maybe we can assume that the more positive the effect score, the more likely the gene is to suppress tumor growth?
  2. gene dependency is the probability that a gene is essential, right? Then can this number only be used to identify genes that promote tumor growth (>0.5), but not genes that inhibit tumor growth?

Thanks

2

“gene essentiality” is trying to capture the notion that a gene is necessary for proliferation. It certainly is reasonable to imagine that there’s an inverse property (genes which when knocked out confer a growth advantage) but I’m not sure what label is most commonly used to describe this.

For example, if one knocks out a tumor suppressor, you might see that’s advantageous to those cells and result in additional proliferation. However, I’ll note that the CRISPR screens generated at the Broad were designed to identify “essential genes” and I cannot speak to their sensitivity to detect genes which promote growth.

I’ve seen remarks discouraging interpreting those genes with a positive gene effect as necessarily being growth advantage. I don’t know that we have sufficient validated examples demonstrating that’s the case.

You also asked:

We know that the more negative the gene effect score, the more likely the gene is to promote tumor growth.

I think that’s not completely accurate. I’d say the more negative the gene effect score, the more likely the gene is essential to growth of that cell line. There are many genes were are essential in all cells types (“pan essential genes”) which is to say essential in both tumors and normal tissue.

Observing a gene is essential doesn’t tell us anything about whether it is specifically related to tumor growth or not.