An epigenetic editor to silence genes

When I was a postdoc at berkeley, my advisor gave me a project to work on. A lab down the road at Stanford working with yeast had done a series of knockouts, precisely eliminating one gene at a time. They reported a number of fatal knockouts in genes that previously had no known function (or a non-necessary function), concluding that the genes presumably had some sort of necessary function for viability (such as housekeeping genes, but not in an obviously necessary way).

Since I'm a DNA nerd I spent time looking at the nature of the genes that were knocked out, and did a bunch of analysis. First I categorized the genes in various ways and didn't see any patterns. But, in the past I'd heard of overlapping genes, and mostly for fun/edification, I taught myself some CS I didn't know and created a data structure and algorithm that allowed me to find all the pairs of 'overlapping' genes (https://www.nature.com/articles/s41576-021-00417-w) in the yeast genome and found that for every single gene they had newly identified as necessary, it overlapped an already known housekeeping gene (overlapping genes were not, in the early 2000s, widely appreciated).

I spoke to my advisor and told him that I reasoned that the authors had made a mistake: in every case where they identified the gene as being necessary, they had accidentally disabled a housekeeping gene that was already known to be necessary for viability, and incorrectly concluded that the gene they knocked out intentionally was the cause of loss of viability. I hope that makes sense- it's basically a "false positive" that had another explanation: they had knocked out two genes when they thought they had knocked out one.

My advisor agreed and I sent my results to the authors, who never responded. Their subsequent paper explicitly mentioned my observations, without crediting me.

Since then, I've come to believe that much of what we believe about genotypes and phenotypes, even in cases where the outcome seems quite straightforward, linear, and single-cause, is instead an error of preconcieved assumptions. It greatly reduced my confidence in geneticists (I'm a biophysicist- highly quantitative, interested in the molecular cause-and-effect) ability to make strong statements.

I don't work in this area any more (it's more lucrative to move data for biologists than it is to be a biologist) but I strongly suspect that I if enabled my OCD bit and spent more than a month analyzing MSTN knockouts, I would find that something which is "self-evidently obvious in MSTN knockouts" is in fact much more complex and subtle than the narrative in the literature.