Before cancer kills, it cheats
news.harvard.edu
news.harvard.edu
I used to work at a lab as an undergrad that worked on extremely similar topics, though the cancer link was more implied: http://research.fhcrc.org/shou/en.html
We had four yeast systems: Two cooperative yeast systems (exchanged metabolites the other needed to survive), and two cheater systems (needed those same metabolites to survive, but just took from the system and never gave back).
One experiment I assisted with was basically putting them all together at differing ratios and seeing what happened over time - the different systems released different fluorescent proteins, so we could measure their relative populations.
Common sense might dictate that the cheaters would rapidly outgrow the cooperators and eventually crash the whole system, but that actually didn't happen until very high ratios of cheaters.
The cooperative yeast systems somehow managed to extinguish the cheaters pretty consistently (all the experiments were repeated and done in triplicate). This happened even at fairly absurd ratios like 3:1 cheaters:cooperators. Unfortunately, I left before any of the follow-up work to determine why that was happening.
All that to say, it's fascinating how much cooperative systems really are favored in nature. Glad to see someone explicitly applying this idea to cancer.
Cheater mutants have been widely observed and fall into two categories: obligatory and opportunistic cheaters. The obligatory cheaters refuse to participate in the stalk, while the opportunistic cheats preferentially segregate themselves into the fruit. It's easy to see how these behaviors might be beneficial. There's a double edge to obligatory cheating, though: in they cannot undergo sporulation.
Look at the functional call stack. Its a slice of the linear address space that’s actively being used and it is strikingly similar to the promoter regions of the genome. Like some pathogens, hackers randomly try to hit that sweet spot that will jump to the special program.
Operating systems are like simple biological organisms that evolve in response to various influences (cybersecurity it seems these days). My favorite is OpenBSD, they invent the cybersecurity. They were one of the first to implement address randomization for kernel and library loading. This is actually strikingly similar to what goes on with the genome with retrotransposons during early embryology stages.
People really like to repeatmask (aka SINE/introns) genomes because it makes BLAST go faster. Ironically, arithmetic/statistical compression algorithms work really well with repeats (I don’t think this compressed state can be examined, but compression algorithms are seriously under appreciated).
Malicious hackers are on the one side very smart (the cancer genesis process is not smart at all), but on the other side they make poor use of evolutionary strategies, mainly because human hackers can't easily spawn "an army of clones of their mind" (cancer otoh is all about evolutionary strategy).
Since the information about "what mutating a codon in a gene does" is not really available to "the attacker", it's as if the whole "address space" is already randomized by default...
The "cancer problems" sound like it's more like being attacked by a really really really dumb hacker, who has access to an utterly insane amount (relatively speaking) of computing power!
(...kind of hard to defend against such an attacker unless you invent modern crypto stuff like hash-signatures, blockchains etc. And our "DNA computer" is not at all designed for these types of math, recoding it to do this would probably be harder than developing mind upload capabilities or superhuman-AI :P Hence not likely to happen.)
As for cryptography, I would argue the biology contains hash signatures and other configurations common to cybersecurity(blockchain turns one problem into many IMO). While there may not be asymmetric cryptography, there is also no pseudorandom business.
There's also a level of programmatic structure in terms of transcription promotors and gotos, and probably some level of conditionals. DNA is also special in that it can be read forwards or backwards, and the base pairing means you always have two potentially transcriptable sequences, each starting from one of 3 offsets. Imagine if busybox implemented all of coreutils in one binary by shifting the word alignment 1 bit to produce a completely different program.
Finally, at the intracellular level, we have a whole host of message passing between cells using hormones and other signaling molecules, and organs like the pancreas that try to regulate a system using them.
As far as 'computation' I don't think the system is Turing complete, and the systems aren't designed to answer decision problems, just construct things, possibly in response to an environmental factor.
Doubtful. You can to computing with DNA http://dna.caltech.edu but it is still debated if the actual role of DNA is being a code (some recently weasel out form that stance by saying it's an “app”, like there's a difference, machine code being a code). Coding theory applied to DNA yields inconclusive results. Galois theory usually has power over any kind of information encoding, cryptographic, computable or not. One constructed one for DNA convinced mathematicians it's not the way to go at all. If it's computation it's nothing like what we mean by any model computation, you may as well say it's magic instead of making stretched analogies.
About OpenBSD and evolution, that's so fetch... these folk invented the attack in the first place, not evolved a response to some market force in the early oughties. Broadly I don't think designed systems are in business of evolving, otherwise living organisms could perhaps evolved electro-hydrostatic instead of hydraulic power system with a pump being a single point of failure.
A cell must sense and respond to its environment. One aspect of this is regulation of gene activity by interactions of for example, transcription factors. When interaction types can be well enough approximated as either of inhibition or activation, you can model them with boolean networks and when levels matter, some have found moderate success with recurrent neural networks under a restriction on what NN nodes represent, to ease interpretability.
That is not the same thing as saying your genome unrolls neural networks. What it is actually saying is that the complexity of the best performing neural network indicates the richness of the underlying computation, and the predictive accuracy of the model captures the functional equivalence of the respective computations (in the cell and in the neural network). The learned model is the instance and neural networks are the class, it is a mistake to place emphasis on neural network, they are merely a way of packaging the computational model of interest.
A cursory review here: https://www.sciencedirect.com/science/article/pii/S001048251...
If you have time for videos: https://www.youtube.com/watch?v=vA9D727AGHI , https://www.youtube.com/watch?v=ZX-GrO2qXQM
More relevant to this article is the study of cancer within the framework of evolutionary game theory, as done here: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC2768082/
There is a precise correspondence between that formulation and one done in terms of online sequential prediction or the so called experts algorithm. All of these are models and still quite limited at that but biology is complex and the future lies in the development and leveraging of these rich connections.
In biology there is 2 factor authentication to protect against parasitic spoofing, intrusion detection systems, protected kernel memory, memory hierarchy, policy based permission. It a very nice operating system IMO.
There are things that concern our electronics that do not matter for biology and there are many things biology must allow for that our hardware cannot. Tying things down to the peculiarities of our systems too narrows the scope of applicable models and understanding. Sometimes it is more useful to think in terms of stochastic differential equations than in terms of operating systems.
The properties of the genome are similar to the storage memory in a computer in that malicious code is inserted into the path of the biological interpreter. Early during human embryonic development (and only during this time for the human) you have a sudden increase of retrotransposon activity. The DNA itself is like stored polymorphic computer virus with a compiler and code to interface with the environment. It’s very unusual the timing and genes carried around by this circular process are vital for life. That is my basis for the comparison to ASLR. There also exists 2 factor authentication, signature based intrusion detection, intrusion prevention systems.
Here is an interesting paper on an organism that has high rates of this DNA recombination..read the motivation section and you decide if this is computation.
https://ac.els-cdn.com/S0166218X09002534/1-s2.0-S0166218X090...
I once attended a lecture at a physics conference by a guy who basically used gzip to do sequence alignment (using compressibility as a way to measure Kolmogorov complexity). Might be interesting to you. The title was: Blasting and Zipping: Sequence Alignment and Mutual Information
I may be misunderstanding what the author proposes, but if the threat of cancer was reduced to the physiological impacts of dormant tumors, I'm pretty sure we would consider that a win.
Regardless of whether you think the specifics presented in this article (e.g., NSAIDS-esophageal cancer) are right, it’s great to see this kind of philosophical approach to biomedical research. We need more people like this professor.
That she started with a plant and that made her think about humans was a great insight. She rightly recognizes the key role of cell autonomous vs. cell non-autonomous processes in how multicellular organisms work.
That she is a woman and someone not at an Ivy League school is also very refreshing to see getting publicized by Harvard. There are so many great scientists aground the world doing great stuff. Let’s keep finding them and reporting on them!
Please media, more of these types of stories and less clickbait please! I loved this article. My faith in humanity was restored just a touch here.
This is an odd comment: This realisation is quite literally at the core of cancer research, and every researcher in the field recognises this. Heck, it’s implied if not mentioned outright in the seminal Hallmarks of Cancer paper [1] that every cancer researcher (and probably generally most human biologist) has read.
Don’t get me wrong: it’s a well-written article explaining the biology to a lay audience. But it doesn’t hold any striking new insights. All of this is not merely known but in fact fundamental knowledge.
This "article" in the Harvard Gazette is basically a press release about an invited lecture that happened at Harvard. If it manages to clearly and correctly explain any science, then it's already going above and beyond its main purpose.
Edit:
* Here is Athena Aktipis' lab site: http://www.aktipislab.org
* Here is a 2015 NYT article about her work that goes into more detail than this press release: https://www.nytimes.com/2015/07/28/science/cellular-cheaters...
* This is the scientific article: http://rstb.royalsocietypublishing.org/content/370/1673/2014...
A lot of "fundamental knowledge" in medicine is wrong. The genetic theory of cancer has always been a dead-end, because it reverses cause and effect. The cancer industry has encouraged "kill all the cancer cells" treatments because it's convenient for their business model. The cancer treatment situation is similar to how the use of bloodletting persisted for hundreds of years [0], in spite of the low rates of success.
[0] http://www.nbc.com/saturday-night-live/video/theodoric-of-yo...
There's no need for new insights, the old insight needs to be fully developed. Your link has a subsection:
https://en.wikipedia.org/wiki/The_Hallmarks_of_Cancer#Deregu...
Warburg's nobel prize was in 1931; we're finally realizing he was probably right all along.
(minor edit)
Most therapeutics today are developed knowing that they will not have a substantially more beneficial effect than current therapies. We still develop these because even if they are not better they could reduce adverse events caused by our current best meds.
As far as I can tell in the last 80 years the Warburg effect has led PET scanning a viable way to detect the progression of many cancers.
To suggest that increased research into the Warburg effect is going to be the thing that leads to an explosion in novel therapeutics for cancer is wishful thinking, as far as I'm concerned. Cancer is complex, and while we could stand to learn more about metabolic changes in cancer, there's not going to be a magic pill for this one.
I like to say that sometimes doctors do good work, and sometimes they make work for themselves. Please take this response as a gentle and well-intended effort to protest the standard practices that hurt patients by 'missing the forest for the trees', such as I've observed in my own family (who themselves are employed in the conventional approach to medicine), and my girlfriend (who is being harmed by allopathic psychiatry).
Did you know that the term 'allopathy' was coined by a homeopath? At the time bloodletting, calomel, and blister agents were the standard of care. Homeopathic medicine held that the body can fix itself when properly supported (diet, clean environment, etc). Allopathic medical practitioners engage in the heroic struggle against disease [0].
[0] https://en.wikipedia.org/wiki/Heroic_medicine
Allopathic medicine never went away, it just changed form. While conventional medical practitioners are doing much better than 200 years ago (w.r.t. not killing their patients), from my perspective, conventional psychiatry and conventional cancer treatments leave much to be desired.
> To suggest that increased research into the Warburg effect is going to be the thing that leads to an explosion in novel therapeutics for cancer is wishful thinking, as far as I'm concerned.
I have some great anecdotes about cheap metabolic therapies. But they're just my non-rigorous anecdotes... The one person who I'd most like to experiment on, trusts her doctors because they wear the halo of 'conventionality'.
> Cancer is complex, and while we could stand to learn more about metabolic changes in cancer, there's not going to be a magic pill for this one.
Consider the possibility that cancer is much simpler than the geneticists have led us to believe. Sometimes 'we must unlearn, what we have learned'.
Edit: google scholar turns up some papers about 'spontaneous remission' of cancer. To figure out how to induce spontaneous remission of cancer cells would be a revolutionary advance.
There's a lot of work being done on that front, primarily in the form of inducing an immune response to cancer.
This work is being done in "allopathic" medical schools and hospitals and in biology labs that are the source of most allopathic medicine.
If this work succeeds, it might lead to a cancer vaccine: the kind of treatment sometimes opposed by homeopaths and naturopaths.
'Allopath' is considered a derogatory term by the modern doctors who are aware of its origin.
> If this work succeeds, it might lead to a cancer vaccine: the kind of treatment sometimes opposed by homeopaths and naturopaths.
The core of the resistance to vaccines is an opposition to the idea that a disease's context doesn't matter. Improved sanitation, refrigeration, insights into adequate nutrition, and other improvements in technology have done just as much as vaccination to vanquish the diseases of antiquity.
I don't understand the relevance. Will sanitation, refrigeration, or nutrition protect you from polio, measles, or other diseases commonly vaccinated against?
Homeopathy is pure, unadulterated garbage, and proof that people will repackage old ideas such as sympathetic magic in new forms. You might as well rub a toad on your head. Homeopathy capitalizes on the same simple fact that most medical scams do, which is that most illness will indeed resolve with time. If you then take credit with placebo interventions for the body’s natural processes, you can fool a lot of people.
It’s funny, until someone decides to forgo treatment for a curable illness in favor of that magical thinking.
I have been tangentially involved with pain researchers that teamed up with chiropractors to study some of their maneuvers in a rigorous manner.
Thanks for the suggestion!
“Homeopathy is a pseudoscience” [1].
[1] https://en.m.wikipedia.org/wiki/Homeopathy second sentence
But the fact is there have been great strides in the survival rate of many kinds of cancer over the past 40 years, even though some are more stubborn than others. It hasn't been dramatic advances for most of it; most of the gains have been slow, incremental improvements that add up over decades.
> most of the gains have been slow, incremental improvements that add up over decades.
I read something 1 to 2 years ago about how breast surgeons aren't recommending prophylactic mastectomies anymore. The message hasn't spread very far. The official approach to breast health is at odds with what's actually been discovered by physiologists.
I know a lady who had "colon cancer". I think her actual problem was emotional stress about her son... But she had her colon shortened, and is probably counted as a success by the official medical statisticians.
Rings a bell
Thinking beyond the snark for a bit, though, locusts might be an example of a species-level cancer cheating the “rules” of a habitat.
Humans? We just throw the rules away altogether and have historically written new ones that only apply to us. The difference between us and locusts, though, is that many of us fight back against our own short term interests and argue “hey, we’re a tumor right now. We should consider not being one of those.”
Alas, we're the only sentient beings in this comparison, which makes tumors bad and universe a resource, by our standards - i.e. the only ones that matter.