In bacteria at least it's well-established that natural selection favours a smaller genome -- hence why antibiotic resistance will be dropped if the selective pressure is removed.
Unfortunately, it is not dropped completely, just reduced in the bacterial population. It would seem a simple idea to just rotate antibiotics (use for a few years then put on the shelf for a few years), but once a bacterial pathogen acquires resistance it never really loses it and once you start using the the antibiotic again the resistance roars back very quickly.
This line of thinking really highlights how building models that predict phenotype from genotype assuming linear and independent genetic contributions, though useful for predicting something like height [0] [1], is not going to generalize sufficiently to solve the lofty goal of personalized precision medicine even with an explosion in data.
Biologically, we know tons and tons about the interconnected biochemical pathways that cause a lot of activity in the cell. It's impressive that statistical modeling approaches that ignore this knowledge earned by thousands of humanyears worth of work can work so well in some domains, but I'm wary of the possibility of academic communities isolating themselves from empiricism and continuing to make mathematically-convenient assumptions, which in the long-run helps no one but themselves.
I don't think this is ignored by the scientific community at all. For example, Eric Lander's group has a pretty nice treatment of why additive models can't explain heritability [1]. And this is coming from the same Eric Lander who has done some of the key work in complex disease / additive genetic modeling.
Imagine you're in a world where you are handed down some knowledge there are Newtonian models, and more advanced models that use relativity, but you don't know what the constants are yet. We're still looking for the Newtonian constants. We know there is much more out there once we understand the basics. But the basics are still worth understanding.
Hypothesis: it’s not so much the content of the genes that’s brittle, as that there are very few proteins as “smart” in how they walk the genome as CRISPR is (and even it’s no Einstein.)
There are likely cellular genetic repair mechanisms—especially in “small” genebases that don’t use fancy tricks like methylation/acetylation—which “hardcode” assumptions about gene number, gene size, positions of important proteins, etc.
Just because DNA itself is a robust data-structure (with features like start-stop tags, commented sections, etc.) doesn’t mean that all the algorithms (proteins) that have evolved to utilize that data-structure query or mutate it in equally-robust ways.
Imagine a codebase started off by a brilliant and strict engineer... and then followed on by a million forks-of-forks-of-forks, all done by average Joes just “scratching their own itch.” That’s the type of software encoded in 99% of the DNA on Earth. The codebases where “every line counts” and has been carefully selected for are far fewer—mitochondria, certainly, but i hesitate to name any other clear candidate.
The proteins that organize DNA, called histones, can be chemically modified in a variety of ways. The modifications have the effect of turning on/off stretches of the DNA simply by modifying its physical structure (the way the dna is arranged and packed around the histones). There are tons more mods that can happen which have a variety of effects (for example, tune the amount that a gene is expressed). This is called epigenetics if you want to google more.
If you've heard of those studies where if your grandparents experienced a famine it makes some difference in your genetics, one hypothesis is that its caused by this.
> Imagine a codebase started off by a brilliant and strict engineer... and then followed on by a million forks-of-forks-of-forks, all done by average Joes just “scratching their own itch.”
I would be careful with that example, before you start to sound suspiciously creationist. At least, the Christians I know that deny evolution tend claim everything was created perfectly and since has only slowly degenerated over time.
As you meet more Christians you will find more variation in what they believe. All should consider themselves "Creationist", but there are MANY competing ideas, from Ken Ham "5000 years old or bust", to Catholic "could be billions of years", and a lot of version in between. For instance, a friend of mine is a biology professor at the local university, his PHD is in fish evolution, and is a very conservative Christian.
Francis Collins is a well documented example of a Christian who very much contrasts Ken Ham as the current head of the NIH and director of the Human Genome Project.
https://en.wikipedia.org/wiki/Francis_Collins
It comes down to what you hold to be true. If you believe Genesis is historical fact, then the interpretation of the science facts are wrong. If you believe the Bible is true and that the interpretation of science facts are true, then you see Genesis in a figurative interpretation. If you reject the Bible, then you will see the narrative of the Bible as false and the scientific interpretation as true.
you kind of need to follow your own advice there. "all should consider themselves Creationist" would be considered pretty silly, or at least absolutely debatable, in the church I grew up in (a very unremarkable suburban Presbyterianism). the Bible itself contains two contradictory creation stories.
I think the overwhelming majority of Christians fit into the "the two mutually contradictory creation stories in Genesis could not be taken literally even if they didn't contradict each other" camp that I'm most familiar with, or, at most, are nearer the "could be billions of years" end of your spectrum.
edit: after a little Googling, it looks like 51% of US Christians accept a "but it was God's idea" version of evolution, i.e., evolution is how it works, God set evolution in motion. and we probably all know that the US is the only technologically advanced nation with such a high percentage of evolution skeptics. but if it's about all Christians, worldwide, evolution skeptics could be the majority, given that technologically non-advanced countries may contain the majority of the world's Christians.
The "brilliant and strict engineer" in this case was the extremely high selective pressure among the first replicators, where only those who could reproduce relatively robustly survived. DNA was a feature created as part of an intense arms-race for domination of the primordial soup; it was well-engineered in the same way most (retrospectively) successful wartime weapons programs are well-engineered.
The features built on top of DNA, however, were built under environments much closer to "peacetime" than "wartime." (Well, maybe not all of them—the "keeping out parasites" logic is also pretty core. But once you can do that, and especially once you have cells specialized for that role in amongst a multicellular body, the other cells can start to evolve some pretty "bureaucratic" notions about how to do things that would never fly if they were directly exposed to a hazardous, resource-scarce environment.)
1. There was a lot of competitive selection pressure because nothing was complex enough to have any defence against anything else eating it: that is, complex enough to have any homeostatic abilities to continue to survive when a neighbouring organism dumps metabolic waste products into the environment that make it less hospitable; or complex enough to survive when any neighbouring organism gets close enough to just start using its constituent proteins as reactants in its own metabolism.
When you haven't yet evolved adaptations like "strong cell walls" or "the ability to recognize and excrete substances antithetical to your metabolic processes" or "the ability to store metabolic energy for times of resource scarcity", you have to spend a lot more of your time as an organism either moving away from danger, or reproducing away from danger.
And, in environments that cycle between resource-rich and resource-poor, the "reproducing away from danger" strategy is highly preferred: you can always shoot off a germ cell specialized to survive under such resource-poor conditions—i.e. a spore—until it enters a richer/safer environment; and so, when you're in a rich/safe environment, it pays off to have another morphogenic stage specialized toward "thriving" rather than "surviving." (If you're wondering why you'd do this rather than just "turn into" a spore, and then "turn back into" an adult organism, like some bacteria essentially do when dried out—well, that turns out to require more genetic code, especially when you're already doing mitosis anyway.)
2. Things that didn't have robust reproduction mechanisms couldn't evolve these other complex adaptations (cell walls, homeostasis, etc.), because every reproductive step would introduce too many errors for the adaptive "signal" to survive well-enough to spread through the population.
Before any other adaptation, the most important and urgent adaptation an organism that relies upon reproduction can evolve is to increase the robustness of its copying logic to the point that other traits can be constructed cheaply, through focused, linear adaptation rather than random chance.
Or: the probability of any random complex adaptation evolving without DNA is very low. That includes DNA itself, of course. But the probability of any random complex adaptation evolving given DNA is much higher. So you'd expect that, of all the evolutionary lines that survived, they all ended up adapting to use DNA first—because the ones that did some other adaptation first would still have to also evolve DNA [through random chance] to preserve that adaptation for any useful number of generations. Organisms with DNA will strictly outcompete those without, because DNA makes you able to get to any other adaptation much faster—and keep it once you have it, and build on it with further adaptations, until you've got a complex feature.
No that is not true, there are genes that does not contain protin coding segments.
There is no clear definition where a gene starts and ends. Some define it as a segment that codes for a specific feature but that is a recursive definition.
19 comments
[ 2.4 ms ] story [ 54.8 ms ] threadBiologically, we know tons and tons about the interconnected biochemical pathways that cause a lot of activity in the cell. It's impressive that statistical modeling approaches that ignore this knowledge earned by thousands of humanyears worth of work can work so well in some domains, but I'm wary of the possibility of academic communities isolating themselves from empiricism and continuing to make mathematically-convenient assumptions, which in the long-run helps no one but themselves.
[0] https://news.ycombinator.com/item?id=16392877 [1] https://www.biorxiv.org/content/early/2017/09/19/190124
Imagine you're in a world where you are handed down some knowledge there are Newtonian models, and more advanced models that use relativity, but you don't know what the constants are yet. We're still looking for the Newtonian constants. We know there is much more out there once we understand the basics. But the basics are still worth understanding.
1 = http://www.pnas.org/content/109/4/1193
There are likely cellular genetic repair mechanisms—especially in “small” genebases that don’t use fancy tricks like methylation/acetylation—which “hardcode” assumptions about gene number, gene size, positions of important proteins, etc.
Just because DNA itself is a robust data-structure (with features like start-stop tags, commented sections, etc.) doesn’t mean that all the algorithms (proteins) that have evolved to utilize that data-structure query or mutate it in equally-robust ways.
Imagine a codebase started off by a brilliant and strict engineer... and then followed on by a million forks-of-forks-of-forks, all done by average Joes just “scratching their own itch.” That’s the type of software encoded in 99% of the DNA on Earth. The codebases where “every line counts” and has been carefully selected for are far fewer—mitochondria, certainly, but i hesitate to name any other clear candidate.
If you've heard of those studies where if your grandparents experienced a famine it makes some difference in your genetics, one hypothesis is that its caused by this.
I would be careful with that example, before you start to sound suspiciously creationist. At least, the Christians I know that deny evolution tend claim everything was created perfectly and since has only slowly degenerated over time.
I think the overwhelming majority of Christians fit into the "the two mutually contradictory creation stories in Genesis could not be taken literally even if they didn't contradict each other" camp that I'm most familiar with, or, at most, are nearer the "could be billions of years" end of your spectrum.
edit: after a little Googling, it looks like 51% of US Christians accept a "but it was God's idea" version of evolution, i.e., evolution is how it works, God set evolution in motion. and we probably all know that the US is the only technologically advanced nation with such a high percentage of evolution skeptics. but if it's about all Christians, worldwide, evolution skeptics could be the majority, given that technologically non-advanced countries may contain the majority of the world's Christians.
The features built on top of DNA, however, were built under environments much closer to "peacetime" than "wartime." (Well, maybe not all of them—the "keeping out parasites" logic is also pretty core. But once you can do that, and especially once you have cells specialized for that role in amongst a multicellular body, the other cells can start to evolve some pretty "bureaucratic" notions about how to do things that would never fly if they were directly exposed to a hazardous, resource-scarce environment.)
Maybe the “codebase” was simpler at the time. But it worked far worse. A good software analogy might be today’s Word vs the first notepad.
When you haven't yet evolved adaptations like "strong cell walls" or "the ability to recognize and excrete substances antithetical to your metabolic processes" or "the ability to store metabolic energy for times of resource scarcity", you have to spend a lot more of your time as an organism either moving away from danger, or reproducing away from danger.
And, in environments that cycle between resource-rich and resource-poor, the "reproducing away from danger" strategy is highly preferred: you can always shoot off a germ cell specialized to survive under such resource-poor conditions—i.e. a spore—until it enters a richer/safer environment; and so, when you're in a rich/safe environment, it pays off to have another morphogenic stage specialized toward "thriving" rather than "surviving." (If you're wondering why you'd do this rather than just "turn into" a spore, and then "turn back into" an adult organism, like some bacteria essentially do when dried out—well, that turns out to require more genetic code, especially when you're already doing mitosis anyway.)
2. Things that didn't have robust reproduction mechanisms couldn't evolve these other complex adaptations (cell walls, homeostasis, etc.), because every reproductive step would introduce too many errors for the adaptive "signal" to survive well-enough to spread through the population.
Before any other adaptation, the most important and urgent adaptation an organism that relies upon reproduction can evolve is to increase the robustness of its copying logic to the point that other traits can be constructed cheaply, through focused, linear adaptation rather than random chance.
Or: the probability of any random complex adaptation evolving without DNA is very low. That includes DNA itself, of course. But the probability of any random complex adaptation evolving given DNA is much higher. So you'd expect that, of all the evolutionary lines that survived, they all ended up adapting to use DNA first—because the ones that did some other adaptation first would still have to also evolve DNA [through random chance] to preserve that adaptation for any useful number of generations. Organisms with DNA will strictly outcompete those without, because DNA makes you able to get to any other adaptation much faster—and keep it once you have it, and build on it with further adaptations, until you've got a complex feature.
There is no clear definition where a gene starts and ends. Some define it as a segment that codes for a specific feature but that is a recursive definition.
Here is an article about what I mean:
https://www.gmwatch.org/en/news/archive/2000/8975-genetic-de...