Hacker News Reader

Comment by 0xbadf00d | Hacker News Reader

I would agree that this is a significant milestone - however I would have thought that there's still a large amount of randomness in protein expression meaning you would still be painting in "broad strokes" as opposed to granular detail of traits.

View on HN

I don't really know what you mean. Genes encode to proteins, if a disorder is genetic it generally means one or more genes are incorrect, encoding to the wrong proteins. If genetic expression was significantly random, how would we identify genetic disorders to begin with? How would they be hereditary?

Not an expert, but there is a lot that is probabilities. E.g. many genes play a part in disease, so changing one set may get you a risk reduction but not full prevention.

The other issue is epigenetics. Gene expression is regulated by various factors and that gene expression can be passed between generations.

https://en.wikipedia.org/wiki/Epigenetics

Well, broadly speaking, the idea is that there may be a subtle underlying mechanism in gene expression that we are unaware of.

So maybe you'll be able to remove a gene that could cause albinism in an individual, then 25 years later discover that the individual has become sterile.

It may seem otherwise but we still know very little on how a set of genes encodes information of how an organism will develop.

While I'm sure some variant of that will occur at some point, many genetic disorders are relatively simple and well understood. Even if we are still learning, there are dozens of disorders we could eliminate if we had robust genetic editing capability with pretty good odds of success.

"... many genetic disorders are relatively simple and well understood."

Are they, really?

As far as I know (and I am not a specialist, so someone more qualified may correct me), gene transcription and translation is far from well understood.

It's as if we don't have access to the source code for the compiler, only the machine code and the syntax. You can draw correlations and infer causality from your changes, but you can't really be sure.

Not an expert either: Even when we know how it happens, we don't know when it happens i.e. when certain chunks of replication is triggered and when it's not. Environmental factor, hormones, diet, mix of them all and much more etc.

We understand the compiler pretty well. You can grab a Bio textbook and find incredibly detailed diagrams of how the genetic transcription process works. It is true that we don't understand every factor, but it's likely the most significant ones are the ones we understand best.

It depends.

At one end of the spectrum, Sickle Cell Anemia is caused by a single point mutation. The sixth amino acid in the hemoglobin beta chain should be glutamic acid, but it's been replaced by a valine. These have different charges, which is enough to warp the cell's shape and impair its function.

At the the other extreme, a lot of different gene variants have been associated with autism. However, these tend not to replicate very well--the genes identified in study A don't show up in study B and vice versa, and there are other possible mechanisms, like copy number variants.

CRISPR could potentially fix the former, but the latter is way beyond our current understanding.

Yep. As was in the news a couple of days ago, sickle cell disease can be effectively 'cured' by installing an updated Hemoglobin protein [1]. (Some) Breast cancer associated genes have pretty well understood issues that can be effectively remedied with gene therapies [2]. And mutations in and around the oncogene P53 have pretty well understood effects - all the way through public policy with regards to HPV vaccination [3].

Adding single proteins and flipping single genetic bits in single cell types in single organs is not the kind of thing that generally causes cascade failures. There are always exceptions, but in general biology is pretty robust to such changes. In fact, those are exactly the kinds of changes that occur from generation to generation. And these are precisely the kinds of systems that have been under deep study for the past four or five decades.

[1] https://serotiny.bio/notes/proteins/hbb/ [2] https://serotiny.bio/notes/proteins/brca1/ [3] https://serotiny.bio/notes/proteins/p53/

Thanks for the insightful responses

Well, it's not all one:one though [1]. If you're messing with transcription factor, who knows what the domino effects will be. Tiny changes might end up triggering phenotypes [2] already present that wouldn't have been expressed without those edits.

The idea that there is only about 2% difference between a chimp and man highlights this point.[3]

[1] https://en.wikipedia.org/wiki/Transcription_factor [2] https://en.wikipedia.org/wiki/Phenotype [3] http://discovermagazine.com/2006/apr/chimp-genome

Disclaimer : Not arguing for or against.

Of all the things to modify though, why a transcription factor? Of course there are things in the genome which modifications of would have problematic results. The question is if we should modify those parts we do think would lead to a better outcome, not the ones we know won't.

Reply on news.ycombinator.com