Breakthrough DNA Editor Borne of Bacteria | Hacker News Reader

Breakthrough DNA Editor Borne of Bacteria

quantamagazine.org

113 pointsmr_tyzic39 comments

TL;DR: Some bacteria use their DNA as a database of bad viruses for identification purposes. This way of editing DNA has been replicated artificially and it's called CRISPR. Awesome.

CRISPR is the type of DNA sequence.

Nature is full of interesting goodies. My favorite is the discovery taq polymerase from thermophile bacteria that lives on thermal vents at the bottom of the ocean.

http://en.m.wikipedia.org/wiki/Polymerase_chain_reaction

Another form of code reuse!

I also wonder what amazing things we could find in rare/niche ecosystems giving that in such a short time we've seen things like this and rapamycin.

Interestingly, someone else holds a patent for this.

http://www.technologyreview.com/featuredstory/532796/who-own...

It's not yet clear who will ultimately own this IP, of Doudna, Zhang, Church (in the US, the European patents have other contenders). Last I heard, Doudna and Zhang were set to have dueling lab notebooks. I wouldn't underestimate Church's position either given his record of visionary patents (e.g. Nanopore sequencing). Disclaimer I work in the Church lab.

Maybe you can enlighten me, but how can any of this be patented in the first place?

Because the patent system is outdated. But the supreme court ruled that "synthetic" or "complementary" DNA is eligible for patent protection.

Basically, it's not the discovery of mechanisms, but the unnatural manipulation of them that is patentable.

But some of the rush to the patent office is because it places the onus on challengers; the actual patent takes years to clear, even without complications.

I agree that the patent system is broken. I'm wondering what exactly about the manipulation is patentable given that the manipulations could happen randomly in nature. I was hoping daughart could provide his arguments as someone who works in one of the labs.

I don't see why this is any different than the patenting of chemicals. Lots of chemicals exist in nature, but if you make something "new" why shouldn't that be protected?

Sorry for the delay - there are many aspects of this that can be patented. Consider all the following inventions:

- Alterations of the RNA that guides cutting (natural system uses two RNAs, which researchers fuse and alter for efficiency of the mechanism)

- Alterations to the protein for efficiency and specificity (e.g. making it nick instead of cut both strands)

- Methods/strategies (how to express the various features, tethering multiple together, how to select only cells that have the changes you want, etc.)

- Applications (use of this system to do something novel)

However, the largest and most central is the use of the system itself for something that does not occur in nature and is non-obvious. This is a system bacteria use as an immune system against foreign DNA. That is not at all related to genome editing. So if people have a problem ("How do we engineer the genome of this cell?"), and then they come up with a novel strategy ("Let's combine these pieces together, using Cas9/CRISPR to cut, then introduce another piece of DNA in a certain way to serve as a template that hijacks the endogenous DNA repair mechanisms to introduce engineered changes."), that is pretty clearly a process of invention to me.

You can't just patent something existing in nature, and as you can see from the examples above that's not at all what is happening here. There is tons of innovation that is novel and non-obvious.

This is going a bit off-topic, but the more I read about these new insights into the workings of single-celled life, the more I get the feeling that on the cellular level, we multicellular organisms are like nations. The more sophisticated ones, like us, being like an ethnically homogeneous and very xenophobic society in a totalitarian state with closed borders). Meanwhile, the single-celled ecologies out there function like some kind of anarchic jungle society were everyone is self-reliant.

It kind of makes me wonder what the future of human society holds.

And cancer is an insurrection against population control that is in the interest of the majority.

We are not a "ethnically homogeneous" society, we are a "clone" society.

In an "ethnically homogeneous" society there are small differences, and it's a good idea for the genes of the small differences to try to get more offspring.

In a "clone" society, there are no differences (or they are extremely small), so it doesn't matter who has the offspring, because it's equivalent. So it's easy to get cooperation.

An interesting cases are the hymenopter (ants, bees, ...) because they have a different system to select sex. For a female is better to have a sister than have a daughter, so the better strategy is to help your mother, then it's a good idea to form a hive.

I'd say we're more alike to the Brave New World society - the cells might be created identical, but because of different environmental influences, they develop different functions.

And the cells are commanded to do their birth-given duty. Any cells that choose to act otherwise, or take part in any activity they were not given access to, including unlawfully reproduce, shall be sentenced to suicide. And if not committed in a timely manner, shall be killed.

>no differences (or they are extremely small)

...other than dramatically altered patterns of gene expression, alternative slicing, epigenetic marks, and irreversible differences in phenotype. Genetics is a small part of the picture and muscle cells don't have much in common with sperm or lymphocytes.

>so it doesn't matter who has the offspring

It matters quite a bit! Controlling who reproduces at what rate and to what homeostatic set-point is a pretty big preoccupation of the body's order. You wants billions of reproducing T-cells that recognize your recently acquired flu but don't want the same level of reproduction among T-cells that react with your gut bacteria. You also don't want uncontrolled proliferation of normally stable populations because then you get horrifying diseases like this: http://en.wikipedia.org/wiki/Fibrodysplasia_ossificans_progr...

Like the others stated, our cells are way to polymorphic to "just" be a clone society - although I agree that an ethnically homogeneous society is still more diverse than that.

Spot on with the ant/bee society thing. In the words of E.O. Wilson: "Communism/Socialism. Great idea, wrong species."

An interesting analogy, but the wild variety of gut flora makes me think that our borders are not so closed. Our own homogenous cells are outnumbered (some studies say 10 to 1) [1] and dependent [2] on our vast colonies of bacteria.

[1]: https://en.wikipedia.org/wiki/Human_microbiome

[2]: http://www.nytimes.com/2010/07/13/science/13micro.html?pagew...

That's basically domestication.

Also, the 10:0 number is highly misleading - the number of bacteria in our body are only around 1 to 2 percent of our body mass.

> Barrangou and his colleagues found that the bacteria had stuffed DNA fragments from the two viruses into their spacer

Sounds almost Lamarckian, doesn't it?

Yes, it is. From a text box in the article:

    Did this mean that CRISPR meets the requirements
    for Lamarckian inheritance? “In my humble opinion,
    it does,” said Koonin.

Oh, cool, I hadn't made it quite that far, I've been slowly reading the whole thing. It's fascinating stuff, and the article seems to be nicely written, in a way that doesn't feel dumbed down, but is accessible to someone not familiar with the field.

Note that we do have a whole field of epigenetics which deals with Lamarckian stuff.

Lamarckism isn’t about epigenetics, nor vice-versa. Some people are calling epigenetic inheritance Lamarckism, but many (in my experience, most) colleagues reject this label as inaccurate. For one thing, although epigenetic inheritance on the single-cell level is a simple fact, and amply demonstrated, it hasn’t been shown to contribute to evolution in any way, and trans-generational epigenetic inheritance in multi-cellular organisms is a difficult concept to demonstrate (there are papers which claim that, but these are contentious, since they fail to demonstrate a persistent effect across several generations). The claims of epigenetic-inheritance-as-Lamarckism are on par with saying that a person’s nationality is Lamarckian: If I emigrate and take on a new nationality, there’s a high chance I’ll confer this nationality to my offspring, simply by raising them in that country. Yet no evolutionary biologist would claim that nationality is in any way part of my genetic heritage.

Here’s a more detailed treatment of these claims: http://skeptics.stackexchange.com/a/7338/82

In a way, all mutations in single-celled organisms that are inherited by daughter cells are acquired during the mother’s life-time, so they’d all meet some definition of Lamarckism. CRISPR isn’t that special in this regard.

Many evolutionary biologists also prefer not to call this Lamarckism – not because it would be entirely wrong, but because it is misleading to many lay people, who imagine that this would somehow be in conflict with Darwinian evolutionary theory of evolution by natural selection (it’s a hobby of many creationists to search for press releases mentioning Lamarckism to crow “ha! Darwin is refuted again!”). This choice isn’t a problem in practice because Lamarckism – unlike neo-Darwinism – doesn’t offer explanatory models which are essential to understand and reason about these processes.

> Many evolutionary biologists also prefer not to call this Lamarckism

Maybe that's why I got downvoted for - as it turns out - writing the exact thing the scientist said.

It was just something that popped into my head.

Yeah, don’t understand the downvotes. It’s an interesting question, and the whole subject of Lamarckism is far from trivially obsolete. Richard Dawkins’ The Blind Watchmaker contains a coherent and convincing argument why Lamarckism is actually wrong as an evolutionary explanation (in a nutshell, Lamarckism fundamentally cannot explain increase in complexity, as Darwinian evolution can, and it additionally doesn’t explain how adaptations of complex traits could possibly be acquired from scratch) – and this is a fundamental, epistemological objection which no amount of evidence could overturn. But few people refer to this reasoning when claiming that Lamarckism is “wrong”.

What about this: sequence a tumour, find a distinguishing part of its genome. Make a virus that search and replaces that with self destructing genes. Inject into tumour.

Then what if: 100 years later, this virus mutates and starts injecting self-destruct genes everywhere? Might be a good, fast way to a zombie apocalypse!

Yup. That's the idea. The problem right now is the cell entry. The crispr/cas has to be placed in the target cell somehow, and yes, we do this by, e.g. as a replacement of antimicrobials, using phages (viruses that attack bacteria. The mechanism must be specific and effective. Now, sadly, most cells and bacteria come with different phages/viruses that allow cell entry. And so it's a _lot_ of systems that have to be engineered and tested.

Of course, there is also the danger of cutting the genome in a wrong place with similar sequence. Again, this can cause adverse effects like loss of cell function or cancer.

Still, therapy with gene editing mechanisms are already been tested, albeit using the older zinc-fing-nucleases for treatmant of e.g. HIV. These are currently in phase 1/2 testing.

So yes, this is currently an active topic of research. But there are still lots of problems that make this technology only feasible in 10+ years for patiant treatment imo.

It's already been used in larger animals. http://www.nature.com/nprot/journal/v9/n8/full/nprot.2014.13...

I know there is one paper that describes curing a liver disease (in a mouse I think) using CRISPR-CAS but I can't find the reference right now.

I can't remember the details of this either, but as I recall it was done in liver because when you inject DNA into mice the liver basically filters it all out, so you get very high efficiency in the liver, but no where else!

A big challenge is that cancer cell genomes are not well conserved. A study was done a while back to look at 1000 breast cancer patients. Mutations were rampant with only a few percentage actually sharing common mutations.

Perhaps the point really isn't about having an effective cure for all cancer patients, but just this one.

One of these days, we can detect the cancer cell, and the specific mutation for that person. The detection would take a day or less, and then prime your immune system to recognize and attack it.

So therapies and cures and not like a flu shot, like a jigsaw puzzle, where the last piece for the cure is different for every person (or a few people).

Or, now that we know MS, schizophrenia, and bipolar disorder arise from an activated endogenous retrovirus [1], what say we nip that in the bud.

[1] http://discovermagazine.com/2010/jun/03-the-insanity-virus

This is something that some people has suspected in the field for a long time. I had terrible discussions as we made some software in the field and met people there.

The pure Darwinism evolution dogma has extended a lot because it was an easy answer, like the world being created in 6 days. But it was incomplete in lots of ways.

For example, considering all DNA we did not understand as "junk" DNA was incredible arrogant, when it seems like it is in fact software or other kinds of data.

You have comments on me on hacker news talking about that like 5 years ago or something.

Great job what those researchers have done.

> considering all DNA we did not understand as "junk" DNA was incredible arrogant

I never heard an actual biologist take that view, only popsci talking heads.

I seem to remember getting the impression that many "actual biologists" did refer to it and often think of it as "junk" DNA in the '80s. But that's just an impression since I never bought into the concept it was "junk" and to the extent I paid attention, it was to those trying to find why it was there and what it did.

I agree. Back in the '80s scientists claimed that a large proportion of DNA didn't encode for anything (introns) and therefore was "junk".

Now our understanding is that although DNA might not encode for anything, it still plays a role in DNA expression.

Reply on news.ycombinator.com