undefined | Better HN

0 pointsImaCake5y ago0 comments

Since you are curious I will try to answer with a bit of a tangent :) But, in short, the real problem is the unintended edits that happen at the same time. Because if you mutate a bunch of stuff in a cell, you will probably end up giving the patient cancer. 99% efficiency is good, but not really essential for many gene therapies.

Something Cas9 does really well is find the sequence you program it to look for, and then bind super tightly to it. You can even change your Cas9 so it will no longer break the DNA. Once you do this, Cas9 cannot edit the DNA by itself, but you can add a second enzyme to it that is able to change a single base of DNA from one letter to another. In theory this means you could target a single nucleotide somewhere in the ~3 billion letters of the genome. We call this changed Cas9 "base editor". It is much more refined than the crude DNA breakage of Cas9, since base editor will initiate only one time of edit in the genome (a single letter flipped from A to G or C to T). But it will still have unintended edits! It is pretty trivial to test for these in whole genome sequencing, and it turns out you will get plenty of them.

Another problem, which might seem more important, but I think is a much easier fix in the medium term (~10 years). Is that base editor will actually target a region of about 5-6 bases around the intended target and will happily flip those bases too. So you are unlikely to get just your intended mutation. Some of your cells will also get other mutations in the same spot. The reason this is less concerning in the long term is that I think this can be fixed with optimisation of the base editor protein. The off target effects are much more difficult to prevent.

0 comments

2 comments · 1 top-level

keenmaster5y ago· 1 in thread

Thanks for the explanation.

>the off target effects are much more difficult to prevent

Why is that the case? And are there any promising theories for how we might overcome that problem?

CRISPR has already been approved for some forms of gene therapy. I assume that's because, for people with crippling diseases, the benefits clearly outweigh the risk. But going forward, as we address less crippling diseases, how do you grok the known benefit/unknown costs of using CRISPR for a cure? This will be an even bigger problem once CRISPR is used for more superficial or cosmetic things. We need to create a decision model and quantify the risks as much as possible.

I have a feeling that countries will "compete" on taking risks with CRISPR. For example, China might see CRISPR as a way to increase its dominance on the world stage. If there's a risky way for CRISPR to increase intelligence, they might take the risk, because China has more to gain than America.

ImaCakeOP5y ago

>Why is that the case? And are there any promising theories for how we might overcome that problem?

I am still quite new to CRISPR so I am unsure what people are doing to solve this problem. As for why, I would say that the problem is that, like most proteins, Cas9 just bounces around using brownian motion and there is always just some probability of it being near the wrong piece of DNA at the right time. There's not much that can be done about that! However, we can do something about Cas9's binding affinity to sequences that are similar, but not idetnical, to the target sequence. One could imagine Cas9 binding (with lower affinity) to a sequence of 20 bases that is incorrect on a couple of bases. Indeed, much of the off-target effects come from this. So if one could modify Cas9 to only bind to the exact sequence that should reduce off-target mutations.

j / k navigate · click thread line to collapse