6 ms·

It's possible that it's very difficult to engineer a virus that can do better than anything nature has come up with. Evolution has had billions of years to try

It's possible that it's very difficult to engineer a virus that can do better than anything nature has come up with. Evolution has had billions of years to try all the attack vectors. A worst case 12 Monkeys-style disaster seems unlikely. But I'm ignorant of this topic. Without handwaving or pointing at the worst case, how much should we worry? Is there any evidence that a genetically enhanced virus can be devastating? https://en.wikipedia.org/wiki/Genetically_modified_virus https://en.wikipedia.org/wiki/Genetically_modified_virus doesn't have much.

drug_design_PhC 9y ago
> It's possible that it's very difficult to engineer a virus that can do better than anything nature has come up with As someone finishing a PhD in computer-aided drug discovery, this seems to be the situation in a lot of cases. Nature often finds it trivial to get around drugs and back to the viral behavior it wants. For example, a lab might make an allosteric drug that reduces the activity of a viral enzyme. In response to treatment with the drug, a few copies of the virus emerge with point mutation that makes them more active. Now a patient has two populations of viruses, and the disease no longer cares about the treatment. The surprising thing isn't just that this happens, but that it happens so rapidly, often requiring just a few days or a single mutation. A quick example of the above is the I585T substitution in [1] One big-picture concept is that genes and proteins are more than their immediate sequences/transcripts. On top of their current identity, they all exist on an mutational "landscape". Remember that some codons are redundant, while some amino acid changes would require two or three base mutations. Therefore, the sequence of a virus encodes both its current identity, as well as nearby evolutionary "alternatives". When a recurring threat comes by, the virus might have encoded in its genes a common defense, maybe through a particularly mutagen-exposed gene location or something like that. In the above case, the virus may have had millenia of exposure to weird dietary natural products and whatnot, and encoded easy rate up/down mutations for its essential enzymes, or orthosteric disruptor mutations for its promiscuous binding sites. This sort of fuckery will have to become a consideration in designing the antivirals of the future. And, to toot my own horn: In an evolutionary sense, our fight against viruses is particularly interesting. Now that computers are helping us accelerate drug development, our brains and computers could be considered as part of our immune systems. In the same way that metal weapons made us safe against macroscopic threats (lions etc), computers and public health infrastructure will become our arsenal against microscopic threats. If computer-aided drug design can lead to faster approval of cheap drugs, we can design hundreds of antiviral molecules and chase a virus all over its evolutionary landscape, until it has nowhere left to run. For out-of-field readers -- An anecdote of cancer cells (not viruses) becoming reliant on a drug in [2] [1] https://www.ncbi.nlm.nih.gov/pmc/articles/PMC296079/ https://www.ncbi.nlm.nih.gov/pmc/articles/PMC296079/ [2] http://blogs.discovermagazine.com/d-brief/2017/10/04/cancer-cells-drugs/#.WjrFFtsVhBc http://blogs.discovermagazine.com/d-brief/2017/10/04/cancer-...
tstrimple 9y ago
Your first sentence is true, but ignores the fact that scientists aren't creating viruses from scratch. They are using what nature made and modifying it. Many treatments rely on a viruses ability penetrate cells. Why recreate what already works so well?
ngvrnd 9y ago
I believe there is reason for concern w.r.t. modifications of flu viruses, for one thing. That is, they know what makes flu easily transmissible among humans, and they know a bit about how to alter nontransmissible viruses to be more so. Given the belief that things like the 1918 flu are believed to be viruses that made this jump on their own, it's quite reasonable to worry that we could do it in a lab.
- bartread 9y ago
  Quite. I don't particularly want to find myself living in the midst of a real life remake of The Stand, nor anything close to it[1]. [1] I am clearly not here referring to the supernatural aspects.
  - QAPereo 9y ago
    At least the supernatural aspects would inject some meaning into what would otherwise be our usually farcical behavior.
dragonwriter 9y ago
> It's possible that it's very difficult to engineer a virus that can do better than anything nature has come up with. Evolution has had billions of years to try all the attack vectors. The traits desired in a bioweapon aren't particularly beneficial to the long-term propagation of a viral strain and thus would be selected against, or at least not selected for, by natural processes.
Simon_says 9y ago
Maybe. Evolution selects for virulence, not death toll. After all, your hosts can't spread you if they're dead.
- darpa_escapee 9y ago
  > After all, your hosts can't spread you if they're dead Arguable. Dead bodies can and do spread disease.
  - jacquesm 9y ago
    But typically not for very long.
  - Simon_says 9y ago
    I don't have a source at the moment, but there's evidence that infectious agents become less deadly over time. There seems to be evolutionary pressure to not kill hosts.
- ridgeguy 9y ago
  This effect may not be helpful to us. Humans can suffer very high death rates due to diseases for which we aren't a host. See the Black Death [1]. The host was fleas, riding rodent vectors. [1] https://en.wikipedia.org/wiki/Black_Death https://en.wikipedia.org/wiki/Black_Death
phkahler 9y ago
If I recall correctly, there is one gene from a particularly lethal virus can be transferred to the flu to make an incredibly deadly nightmare. I don't know why anyone would do that, but it's possible. IIRC this has already been done in a lab. Trying to do the same but targeted seems like something certain types of people would try to do and could easily go awry.
emodendroket 9y ago
"It's possible that it's very difficult to engineer corn that can resist pests better than anything nature has come up with. Evolution has had billions of years to try all the attack vectors. Widespread commercial use of genetically modified corn seems unlikely."
mistercow 9y ago
> Evolution has had billions of years As a general rule, if your argument starts with this premise, you're probably heading down the wrong path. Natural evolution is an extremely limited optimization process with a dramatically different loss function than the ones used by humans.
cowsandmilk 9y ago
> It's possible that it's very difficult to engineer a virus that can do better than anything nature has come up with A lot of the debate was started because 2 groups did so in 2011, independently of each other finding mutations that allow the deadly H5N1 virus to transmit via air[1][2]. One group also did work to model the chance of these mutations showing up in nature[3]. The techniques by which to do this are fairly well established. Knowledge of how to do it safely is fairly limited, as is access to animals in which to test these modifications. * [1] https://www.nature.com/articles/nature10831 https://www.nature.com/articles/nature10831 * [2] http://science.sciencemag.org/content/336/6088/1534.long http://science.sciencemag.org/content/336/6088/1534.long * [3] http://science.sciencemag.org/content/336/6088/1541.full http://science.sciencemag.org/content/336/6088/1541.full
scentoni 9y ago
About a hundred years ago, evolution came up with something that killed 50-100 million people. https://en.wikipedia.org/wiki/1918_flu_pandemic https://en.wikipedia.org/wiki/1918_flu_pandemic Smallpox and plague did far worse.

drug_design_PhC 9y ago

> It's possible that it's very difficult to engineer a virus that can do better than anything nature has come up with As someone finishing a PhD in computer-aided drug discovery, this seems to be the situation in a lot of cases. Nature often finds it trivial to get around drugs and back to the viral behavior it wants. For example, a lab might make an allosteric drug that reduces the activity of a viral enzyme. In response to treatment with the drug, a few copies of the virus emerge with point mutation that makes them more active. Now a patient has two populations of viruses, and the disease no longer cares about the treatment. The surprising thing isn't just that this happens, but that it happens so rapidly, often requiring just a few days or a single mutation. A quick example of the above is the I585T substitution in [1] One big-picture concept is that genes and proteins are more than their immediate sequences/transcripts. On top of their current identity, they all exist on an mutational "landscape". Remember that some codons are redundant, while some amino acid changes would require two or three base mutations. Therefore, the sequence of a virus encodes both its current identity, as well as nearby evolutionary "alternatives". When a recurring threat comes by, the virus might have encoded in its genes a common defense, maybe through a particularly mutagen-exposed gene location or something like that. In the above case, the virus may have had millenia of exposure to weird dietary natural products and whatnot, and encoded easy rate up/down mutations for its essential enzymes, or orthosteric disruptor mutations for its promiscuous binding sites. This sort of fuckery will have to become a consideration in designing the antivirals of the future. And, to toot my own horn: In an evolutionary sense, our fight against viruses is particularly interesting. Now that computers are helping us accelerate drug development, our brains and computers could be considered as part of our immune systems. In the same way that metal weapons made us safe against macroscopic threats (lions etc), computers and public health infrastructure will become our arsenal against microscopic threats. If computer-aided drug design can lead to faster approval of cheap drugs, we can design hundreds of antiviral molecules and chase a virus all over its evolutionary landscape, until it has nowhere left to run. For out-of-field readers -- An anecdote of cancer cells (not viruses) becoming reliant on a drug in [2] [1] https://www.ncbi.nlm.nih.gov/pmc/articles/PMC296079/ https://www.ncbi.nlm.nih.gov/pmc/articles/PMC296079/ [2] http://blogs.discovermagazine.com/d-brief/2017/10/04/cancer-cells-drugs/#.WjrFFtsVhBc http://blogs.discovermagazine.com/d-brief/2017/10/04/cancer-...

tstrimple 9y ago

Your first sentence is true, but ignores the fact that scientists aren't creating viruses from scratch. They are using what nature made and modifying it. Many treatments rely on a viruses ability penetrate cells. Why recreate what already works so well?

ngvrnd 9y ago

I believe there is reason for concern w.r.t. modifications of flu viruses, for one thing. That is, they know what makes flu easily transmissible among humans, and they know a bit about how to alter nontransmissible viruses to be more so. Given the belief that things like the 1918 flu are believed to be viruses that made this jump on their own, it's quite reasonable to worry that we could do it in a lab.

bartread 9y ago
Quite. I don't particularly want to find myself living in the midst of a real life remake of The Stand, nor anything close to it[1]. [1] I am clearly not here referring to the supernatural aspects.
- QAPereo 9y ago
  At least the supernatural aspects would inject some meaning into what would otherwise be our usually farcical behavior.

dragonwriter 9y ago

> It's possible that it's very difficult to engineer a virus that can do better than anything nature has come up with. Evolution has had billions of years to try all the attack vectors. The traits desired in a bioweapon aren't particularly beneficial to the long-term propagation of a viral strain and thus would be selected against, or at least not selected for, by natural processes.

Simon_says 9y ago

Maybe. Evolution selects for virulence, not death toll. After all, your hosts can't spread you if they're dead.

darpa_escapee 9y ago
> After all, your hosts can't spread you if they're dead Arguable. Dead bodies can and do spread disease.
- jacquesm 9y ago
  But typically not for very long.
- Simon_says 9y ago
  I don't have a source at the moment, but there's evidence that infectious agents become less deadly over time. There seems to be evolutionary pressure to not kill hosts.
ridgeguy 9y ago
This effect may not be helpful to us. Humans can suffer very high death rates due to diseases for which we aren't a host. See the Black Death [1]. The host was fleas, riding rodent vectors. [1] https://en.wikipedia.org/wiki/Black_Death https://en.wikipedia.org/wiki/Black_Death

phkahler 9y ago

If I recall correctly, there is one gene from a particularly lethal virus can be transferred to the flu to make an incredibly deadly nightmare. I don't know why anyone would do that, but it's possible. IIRC this has already been done in a lab. Trying to do the same but targeted seems like something certain types of people would try to do and could easily go awry.

emodendroket 9y ago

"It's possible that it's very difficult to engineer corn that can resist pests better than anything nature has come up with. Evolution has had billions of years to try all the attack vectors. Widespread commercial use of genetically modified corn seems unlikely."

mistercow 9y ago

> Evolution has had billions of years As a general rule, if your argument starts with this premise, you're probably heading down the wrong path. Natural evolution is an extremely limited optimization process with a dramatically different loss function than the ones used by humans.

cowsandmilk 9y ago

> It's possible that it's very difficult to engineer a virus that can do better than anything nature has come up with A lot of the debate was started because 2 groups did so in 2011, independently of each other finding mutations that allow the deadly H5N1 virus to transmit via air[1][2]. One group also did work to model the chance of these mutations showing up in nature[3]. The techniques by which to do this are fairly well established. Knowledge of how to do it safely is fairly limited, as is access to animals in which to test these modifications. * [1] https://www.nature.com/articles/nature10831 https://www.nature.com/articles/nature10831 * [2] http://science.sciencemag.org/content/336/6088/1534.long http://science.sciencemag.org/content/336/6088/1534.long * [3] http://science.sciencemag.org/content/336/6088/1541.full http://science.sciencemag.org/content/336/6088/1541.full

scentoni 9y ago

About a hundred years ago, evolution came up with something that killed 50-100 million people. https://en.wikipedia.org/wiki/1918_flu_pandemic https://en.wikipedia.org/wiki/1918_flu_pandemic Smallpox and plague did far worse.