With lymphocytes, we already have a template for nanobots to do the job.
Human cells display fragments of proteins within them on their cell walls. Cancer means that cells have a lot of mutated proteins that they would present on their cell walls. Cancer call also shut down the process of protein fragments being shown on the cell walls but that’s also detectable by lymphocytes.
There’s no good way for a cancer to display all the fragments that it’s supposed to display while at the same time not displaying any mutated protein fragments.
Our body does produce the nanobots. But naturally, our immune system isn’t perfect at making the right nanobots. There are clinical attempts to grow the right nanobots in vitro and use them to attack cancer.
In vitro approaches could be improved on computer models that guide the process.
We need one machine learning model that you give a DNA sequence of a cell and that then tells you what protein fragments that cell will display.
We need another machine learning model to tell us how the lympocytes need to be programmed to match recognize a cell that displays those fragments. It’s a simulation task that’s a bit more complex than what alpha fold is doing.
And then we need a good process to grow those lymphocytes in a cost-efficient manner. Likely, something where you synthesise DNA and create a system where that DNA gets used.
The other way to do it is a protein based mechanism that you inject into each cell. If a specific mRNA sequence matches to the mechanism, it causes the cell to self destruct.
Package the mechanism in a virus and inject the virions directly into each tumor. This would cause the cancerous cells to die and the healthy ones to usually survive. (Just injecting a bunch of junk into a healthy cell can make it fail)
Advantage here is it is a general solution. All cancerous cells have an mRNA that is mutated. You can build a different matching sequence for any cancer. So it is a “cure”, there will never be a cancer you can’t treat.
If you could also grow replacement organs—since this battle may scar the patients existing ones and cause other damage—and everything was done by AI systems using high speed robotics that learn from each mistake across all patients (human doctors learn 1 patient at a time until they are forced to retire from aging, AI doctors can learn from a million attempts in parallel) - theoretically the cancer death rate would trend towards zero.
With that said each patient would probably have to live maybe for years in a sealed biolab, surrounded by racks of equipment and plumbing for their external organ support systems etc. It’s easy to imagine a system that would keep someone alive indefinitely if their body is spread apart across a large life support system. It’s harder to imagine putting them back together in a way that won’t have high mortality rates once they leave from unobservable small mistakes.
Note by “easy” I mean with effective narrow AI. What keeps the patient alive for years when human facilities can’t is each domain of the human body (liver, circulatory, hemopoietic, immune, nervous, etc) is monitored by multiple parallel narrow AI, and each piece of hardware has multiple parallel systems. So it’s parallel redundancy: all the hardware for a domain have to fail at the same time and all narrow AIs have to fail to take a corrective action.
So for example for the patients blood chemistry to get outside the narrow ranged associated with long survival, each living artificial liver must fail at the same time, and all monitoring AI systems must fail to observe this, and other systems for related organs must fail also.
It can be made unlikely, where the probabilities work out that a patient on total life support has a longer expected lifespan than a healthy 12 year old human who didn’t age.
There are many methods. I am not aware of which ones could potentially reach every cell type but so long as you have a library of methods capable of carrying a large enough payload, and the scope of the library covers every cell type, then you do have a “universal method”.
I am claiming a generalized solution to cancer that is still customized for every tumor is possible. Are you claiming otherwise and why?
Be rational. You wouldn’t be so defeatist in another field would you?
I am not aware of which ones could potentially reach every cell type but so long as you have a library of methods capable of carrying a large enough payload, and the scope of the library covers every cell type, then you do have a “universal method”.
No, because cancer can mutate to invalidate ways of passing through the cell wall.
I am claiming a generalized solution to cancer that is still customized for every tumor is possible. Are you claiming otherwise and why?
I described a generalized solution in the post you were replying to that’s customized. You described a different one and one that does not actually work as a universal solution.
Such a mutation has to be possible. (in a probabilistic sense—if it requires hundreds of bases of change, and there is not an evolutionary vector pointing that way (microevolution for the cancer) it will not occur in meaningful numbers to matter)
Your solution fails because it’s the “perfect traitor” problem. It is possible for a cancerous cells to display all the right flags the immune system cannot tell it’s part of the tumor.
It sounds to me like you are thinking about this in the abstract without really thinking through the actual biology.
Substances cross the cell wall because cells have mechanisms to transport them through the cell wall. If the proteins necessary for those mechanisms get disjunctional because of mutations, the mechanism stops working.
A cancer cell has no way to stop bunch of random mutations from happening. The idea that it could sounds to me like it misrepresents what cancer is about by a lot.
I would argue every criticism you make can aim at yourself. Obviously viruses have injection mechanisms that can bypass most defenses. If they didn’t virii wouldn’t work.
The rest of it shows a poor understanding of evolutionary algorithms.
Basically, you don’t understand enough biology to see the difference between the two.
Viruses are evolved to interact with relatively stable targets. If we take COVID-19 for example, the ACE2 receptor is used by the virus. If all human cells would stop producing ACE2 receptors the infection doesn’t work.
There are mechanisms to present peptides or protein fragments on the cell wall. Those can break down, but that’s detectable by the immune system because the cell wall look differently.
If they are not broken down they will display any protein fragments that float around in the cell. It’s the nature of cancer that a lot of what floats around within a cell has mutations.
Not having a bunch of random mutations is the one thing that a cancer cell can’t do due to evolutionary pressure.
With lymphocytes, we already have a template for nanobots to do the job.
Human cells display fragments of proteins within them on their cell walls. Cancer means that cells have a lot of mutated proteins that they would present on their cell walls. Cancer call also shut down the process of protein fragments being shown on the cell walls but that’s also detectable by lymphocytes.
There’s no good way for a cancer to display all the fragments that it’s supposed to display while at the same time not displaying any mutated protein fragments.
Now all we need is the nanobots…
Our body does produce the nanobots. But naturally, our immune system isn’t perfect at making the right nanobots. There are clinical attempts to grow the right nanobots in vitro and use them to attack cancer.
In vitro approaches could be improved on computer models that guide the process.
We need one machine learning model that you give a DNA sequence of a cell and that then tells you what protein fragments that cell will display.
We need another machine learning model to tell us how the lympocytes need to be programmed to match recognize a cell that displays those fragments. It’s a simulation task that’s a bit more complex than what alpha fold is doing.
And then we need a good process to grow those lymphocytes in a cost-efficient manner. Likely, something where you synthesise DNA and create a system where that DNA gets used.
The other way to do it is a protein based mechanism that you inject into each cell. If a specific mRNA sequence matches to the mechanism, it causes the cell to self destruct.
Package the mechanism in a virus and inject the virions directly into each tumor. This would cause the cancerous cells to die and the healthy ones to usually survive. (Just injecting a bunch of junk into a healthy cell can make it fail)
Advantage here is it is a general solution. All cancerous cells have an mRNA that is mutated. You can build a different matching sequence for any cancer. So it is a “cure”, there will never be a cancer you can’t treat.
If you could also grow replacement organs—since this battle may scar the patients existing ones and cause other damage—and everything was done by AI systems using high speed robotics that learn from each mistake across all patients (human doctors learn 1 patient at a time until they are forced to retire from aging, AI doctors can learn from a million attempts in parallel) - theoretically the cancer death rate would trend towards zero.
With that said each patient would probably have to live maybe for years in a sealed biolab, surrounded by racks of equipment and plumbing for their external organ support systems etc. It’s easy to imagine a system that would keep someone alive indefinitely if their body is spread apart across a large life support system. It’s harder to imagine putting them back together in a way that won’t have high mortality rates once they leave from unobservable small mistakes.
Note by “easy” I mean with effective narrow AI. What keeps the patient alive for years when human facilities can’t is each domain of the human body (liver, circulatory, hemopoietic, immune, nervous, etc) is monitored by multiple parallel narrow AI, and each piece of hardware has multiple parallel systems. So it’s parallel redundancy: all the hardware for a domain have to fail at the same time and all narrow AIs have to fail to take a corrective action.
So for example for the patients blood chemistry to get outside the narrow ranged associated with long survival, each living artificial liver must fail at the same time, and all monitoring AI systems must fail to observe this, and other systems for related organs must fail also.
It can be made unlikely, where the probabilities work out that a patient on total life support has a longer expected lifespan than a healthy 12 year old human who didn’t age.
There’s no universal method to inject something into each cell.
There are many methods. I am not aware of which ones could potentially reach every cell type but so long as you have a library of methods capable of carrying a large enough payload, and the scope of the library covers every cell type, then you do have a “universal method”.
I am claiming a generalized solution to cancer that is still customized for every tumor is possible. Are you claiming otherwise and why?
Be rational. You wouldn’t be so defeatist in another field would you?
No, because cancer can mutate to invalidate ways of passing through the cell wall.
I described a generalized solution in the post you were replying to that’s customized. You described a different one and one that does not actually work as a universal solution.
Such a mutation has to be possible. (in a probabilistic sense—if it requires hundreds of bases of change, and there is not an evolutionary vector pointing that way (microevolution for the cancer) it will not occur in meaningful numbers to matter)
Your solution fails because it’s the “perfect traitor” problem. It is possible for a cancerous cells to display all the right flags the immune system cannot tell it’s part of the tumor.
It sounds to me like you are thinking about this in the abstract without really thinking through the actual biology.
Substances cross the cell wall because cells have mechanisms to transport them through the cell wall. If the proteins necessary for those mechanisms get disjunctional because of mutations, the mechanism stops working.
A cancer cell has no way to stop bunch of random mutations from happening. The idea that it could sounds to me like it misrepresents what cancer is about by a lot.
I would argue every criticism you make can aim at yourself. Obviously viruses have injection mechanisms that can bypass most defenses. If they didn’t virii wouldn’t work.
The rest of it shows a poor understanding of evolutionary algorithms.
Basically, you don’t understand enough biology to see the difference between the two.
Viruses are evolved to interact with relatively stable targets. If we take COVID-19 for example, the ACE2 receptor is used by the virus. If all human cells would stop producing ACE2 receptors the infection doesn’t work.
There are mechanisms to present peptides or protein fragments on the cell wall. Those can break down, but that’s detectable by the immune system because the cell wall look differently.
If they are not broken down they will display any protein fragments that float around in the cell. It’s the nature of cancer that a lot of what floats around within a cell has mutations.
Not having a bunch of random mutations is the one thing that a cancer cell can’t do due to evolutionary pressure.
The lipid nanoparticle mechanism used by the mRNA vaccines could have had a cancer destroying payload.
As I understand it the lipids merge with cell membranes, it’s receptor independent.
So unless the cancer cell can evolve to not have a cell membrane it’s vulnerable.