Review of 2023 book: How Life Works: A User’s Guide to the New Biology Philip Ball. ISBN9781529095999
Theoretical biologist here. This is an incredibly important book. I just bought it a few minutes ago and so I’m only partway through the beginning, but it’s summarizing everything people from my school of thought (complex adaptive systems theory, multilevel selection models, and so on) have been arguing for two or three decades. It’s a very fast read so far (probably less so if you’re less familiar with the points the author is making), but I really hope that this book has an impact that’s reflective of the timeliness and cohesiveness (as I am reading into what the author is preparing to argue) deserves.
Yeh! Good to see the rusty machine (and self-deprecating) model fading away and being replaced by real appreciation of the true marvels that have emerged over millions of years. (Science’s mechanical models were all so … 18th century!)
(Not so familiar with biology but did enjoy hearing about the tack Lee Cronin’s taken.)
It’s still a rusty machine even if the maths that control it are a bit more complex
Math controls nothing and only models them. It is CRITICALLY important to remember that mathematical models are ONLY models, no matter how closely they match any sampled data.
Everything is math, it’s what controls gravity and cell division and protein folding - there is no god it’s all just math
Of course the model is only a model but the point is you can use the model to predict real world responses, therefore you can test millions of things and do the one which is most likely to work
No, you’ve internalized it in the wrong way. The mathematical model comes AFTER the real world. It is not related to the real world what so ever except in correlation. Correlation is not causation. ESPECIALLY with a human-made MODEL.
If the maths actually drove ANYTHING, you wouldn’t be saying, “most likely” to work…
Yes that’s a model, I’m not saying the model magically controls reality but that the underpinning reality is math - the reason you always have two apples in a bag when you start with one and add another is because of math, the human model of that isn’t controlling it but if we want two apples in a bag and we currently only have one then we can use our model to determine how many apples we need to add into the bag.
The same is true of more complex systems, if we can accurately model the cellar interactions then we can derive solutions in the same way
That’s the thing: It’s NOT math. Math is an expression of relationships. The underpinning of reality IS NOT math. Ever. Math is a simplification. Always.
I think the difference here is between your conception that reality follows a mathematical model while their conception is that mathematical models follow and try to be reflective of reality.
I think their concern is that, if one believes reality follows math, when the model fails to accurately predict something, the person with the model may wonder what’s wrong with reality. If that person believed the model follows reality they would wonder what’s wrong with the model. The latter perspective will yield better results.
It’s the difference between saying “this is how it works” vs “to the best of my knowledge this is how it works.”
That’s ridiculous, if I think that one apple plus one apple is going to result in three apples then I try it and find it’s actually two I’m not going to blame the universe I’m going to know my understanding of arithmetic is flawed.
It’s not the math of the human made model that controls the universe that would be silly. The model is the current best approximation of the actual math that defines the universe.
An accurate model allows you to predict the outcome of events, like we can predict how many apples will be in the bag. With some things it gets very complicated because there’s lots of things and various possible states but we can model that with statistics and calculous and stuff. We can even make a set of all possibile results and use that as a map to tell us if something is possible, how likely it is and what we can do to make it more or less likely - nothing the guy said was controversial, we can map cellar interactions even if that requires using complex multidimensional math (mathematics have had to get used to doing this sort of math a lot recently so I’m sure they’ll manage)
Did you start with the arithmetic that putting one apple in the bag followed by another would leas there being so many, or did you consistently observe that doing so led to there being two apples until your mind learned the math of 1+1=2?
I think this really comes down to your opinion on whether math was created or discovered. Based on your statements so far I’m guessing you believe math was discovered, as there is some mathematical model completely representative of reality. Through observation we can discover mathematic principles to get closer and closer to that model, not that it would necessarily be 100% achieved. I realize that may be putting words in your mouth, but it’s the best argument I can think of to reach your perspective. Is that about right?
What’s the name of the theory the says cells use the genome like a library of tools?
What a dishonest bs. It’s not the scientists who communicate these dumbed down “theories”, it’s journalists and trivial science books and shows.
Makes me loose all respect for the author.
At the university where I studied professors were constantly talking about what we don’t know. Formulated every theory extremely carefully, there was no “it is like that”. What kind of scientists is he talking about?
Yeah, I graduated with my BS in zoology over 20yrs ago and my professors wouldn’t have talked about genetics as a blueprint even back then. My focus was evolutionary biology and the one sentence in the article on the topic made me cringe. I would guess that people who focused in molecular bio probably cringed through the rest of it.
Craig Venter, the infamous head of the Human Genome Project and who created the first “synthetic” cell, has been saying this stuff for years. It’s remarkable how ahead of the times he is, perhaps because he’s not beholden to an academic institution.
He claims that a “tree of life” is fallacious, that there is no junk DNA, and that the bare minimum genes necessary for a living cell still can’t be determined even after decades of research.
I hope that the authors of the new Extended Evolutionary Synthesis will admit the deficiency of outdated assumptions and reject dogmatic approaches to the theory, as implied by the author of the book reviewed in this article.
How could there be no junk DNA? There are plenty of inserted regions of repeating codons, between regions that are read (outside of replication). DNA replicators are very simple machines, they copy until they’re told to stop, I agree that any junk DNA in the human genome has been there for a very long time, but it’s not difficult to find single cell organisms that have introduced previously non-self DNA in their genome. If that DNA isn’t used besides replication then it’s junk is it not?
Also telomeres are pretty synonymous with junk DNA, until they aren’t, or is every shortening of the telomere removing information vital to a cells function?
So I think I can make the claim that I am an expert in this, at least compared to 95%+ of biological researchers. My research foci include epigenetic and emergent interactions like the ones discussed in the article, and although I am not going to back this up by identifying myself, please believe me when I say I’ve written some papers on the topic.
The concept of junk DNA is perhaps the problem here. Obviously there are large swaths of our genome that do not encode anything or have instructions for proteins. However, dismissing all non-coding DNA as “junk” is a critical error.
Your telomeres are a great example. They don’t contain vital information so much as they serve a specific function-- providing a buffer region to be consumed during replication in place of DNA that does contain vital information. Your cells would work less well without telomeres, so calling them junk is inaccurate.
Other examples of important non-coding regions are enhancer and promoter regions. Papers describing the philosophical developments of stochasticity in cellular function note how enhancers are vital for increasing the likelihood of transcription by making it more likely that specific proteins floating in the cellular matrix interact with each other. Promoter regions are something most biologists understand already, so I won’t describe them here (apologies for anyone who needs to go read about them elsewhere!). Some regions also inform the 3D structure of the genome, creating topological associated domains (TADs) that bring regions of interest closer together.
Even the sequences with less obvious non-coding functions often have some emergent effect on cellular function. Transcription occurs in nonsense regions despite no mRNA being created; instead, tiny, transient non-coding RNAs (ncRNAs) are produced. Because RNA can have functional and catalytic properties like proteins, these small RNAs “do jobs” while they exist. The kinds of things they do before being degraded are less defined than the mechanistic models of proteins, but as we understand more stochastic models, we are beginning to understand how they work.
One last type of DNA that we used to consider junk: binding sites for transcription factors, nucleosome remodelers, and other DNA binding proteins. Proteins are getting stuck to DNA all the time, and then doing things while they’re stuck there. Sometimes even just being a place where a nucleosome with a epigenetic flag can camp out and direct other cellular processes is enough to invalidate calling that region “junk”.
Anyway I’m done giving my spiel but the take home message here is that all DNA causes stochastic effects and almost all of it (likely all and we haven’t figured it out yet) serves some function in-context. Calling all DNA that doesn’t encode for a protein “junk” is outdated-- if anything, the protein encoding regions are the boring parts.
Thank you for taking the time to respond, I respect your knowledge and agree with you for the most part. From an evolutionary perspective there’s very little pressure to cull genetic material that does not have a purpose, genome replication is already taking place and takes very little overall energy/time.
There may not be as much useless DNA in the system as previously thought, but not every codon pair has a use. There are undoubtedly identical transcription codes being suppressed in one section of DNA that are active in other regions, and it may have been useful to have that extra region available if pressures ever applied that caused that region to be reactivated, but if mutation occurred and caused that region to no longer have the original blueprint it was coding for, it could theoretically create actual evolutionary pressure to eliminate/suppress that section of the genome, it could be suppressed/inactive harmful DNA, not junk but also not beneficial.
My biggest hang-up on the whole “every codon has a purpose” argument is that it blatantly ignores the evidence occurring so much more frequently at “lower” life forms. Eukaryotic single cell organisms swap DNA rather readily, it’s a much higher risk/reward mechanism of evolution, a lot of that DNA, if it turns out to be beneficial, will be ancillary to the actual genes with benefit. Plants have genomes that vary in length from generation up generation, often times much larger than required, maybe it’s because they chill in the sun all day and are more susceptible to genetic mutation, but just because there’s extra targets for codon swapping, doesn’t mean that DNA is set there with purpose. It just exists. It may have been beneficial at one point, but it’s only there because it isn’t detrimental enough to have selection pressure repercussions. If pressures were high enough they every codon mattered, (or if it were designed intelligently so that every codon mattered) a lot of genomes (I’m not to nervous to claim I believe all genomes) would be shorter due to junk culling, it’s just such a small factor in the schema that it isn’t ever selected against.
I would encourage you to read the linked Science paper and Dan Nichol’s paper, Is the Cell Really a Machine?
You feel that if a codon isn’t meant for something, if it doesn’t have a purpose– then it is junk. This is a mindset that is reflective of the machine model of the cell. We used to expect that each protein was bespoke for a function, each transcript necessary.
The whole paradigm shift at hand is this model falls flat, even for coding regions. I think you’re actually very spot in here with the prokaryotic DNA or the plant genomes (love me some violets for their weird genomes). Some parts of a genome will rapidly change and appear to serve no real purpose, but the next bite is the important one: even if it seems like there isn’t a purpose, like a top-down prescription for functionality, those regions are still doing something while they are present.
For example, some long non-coding regions affect the likelihood that a person will develop Parkinson’s disease, or in the case of plants with various polyploidies, the relative expression of their genes won’t necessarily change, but the absolute expression may.
Basically, you aren’t wrong that these regions dont have a purpose, because no genes have a purpose. The cell isn’t a machine.
Three cheers for Dan Nichol’s paper.
Here’s a direct link to the PDF found on Philpapers.org.The cell isn’t a machine.
What do you mean by this? I feel like you think the meaning is obvious after everything you’ve said, but it’s not.
Even if we accept that everything you said is true, all it means is that the cell is a very, very complex machine. More complex than current models account for. It’s just chemistry, after all. The chemicals behave in predictable fashion or else life wouldn’t be possible at all. Molecules moving around, transforming, causing other molecules to transform, etc, etc, to turn food into shit and babies. You can always use the word “machine” to describe that, no matter how complex it is. Just like the word “algorithm” can be used to describe the function of code no matter how complex it is, whether it’s a simple path finding algorithm, or the newest machine learning one.
But I probably shouldn’t use the word “function” because that implies purpose, and, as you say, no part of the chemistry of life has purpose. I hope you can detect my snark. That’s a pretty lame argument that’s philosophical at best. The purpose of the machinations of the cell is to maintain life and reproduce. No mater how many times you say it, your words won’t change the fact that that is the purpose of the chemistry of life.
You’ve twisted around the word “purpose” in your head until it has no useful meaning. Nonsense. A molecule can many overlapping, hard to discern purposes. That does not mean it doesn’t have a purpose.
When I say “the cell isn’t a machine”, it is in specific reference to the machine model of the cell, which is a previously established conceptual framework in the field of molecular biology. If you want to understand why that model is falling out of favor today, you’re invited to read the article linked by OP and/or the articles I have linked in other comments.
The gist is that the cell is more complicated, flexible, and emergent than any machine has ever been and will be for the foreseeable future, and the idea that we can simply map the functions of each molecule in the cell to get a perfect “circuit diagram” of how everything plays together is defunct.
I don’t have time to mess with this thread any more. You can either accept what myself (an expert in this field), the author of this publication (which happens to be one of the most prestigious journals in the world), and others who do this research daily are saying about this, or you can not. Frankly, if you are an expert also, the field, the research, and the truth barely cares about our opinion-- it certainly doesn’t care about non-expert opinions on the internet.
So, shall we call it “inactive regions” then?
‘Noncoding region’ seems to be the preferred term.
No, because they are anything other than inactive
I’m not an expert on the subject. I can only repeat what Venter said: “the only junk DNA is in my colleagues brains”. He claims that all DNA has function and that it should not be referred to as junk just because we don’t know the function yet.
He needs to look at some plant DNA, there are places with 50 times now DNA codons per cell than Humans have, with many many many times fewer genes.
“If it’s there it must be there for a reason” sounds an awful lot like intelligent design to me, and his putting down his colleges for holding alternative (seemingly more informed than his own) theories doesn’t help my view of him. More codons don’t mean more reason, evolution is not what is most efficient, it’s just what works best at any time. It’s also full of cross contamination at the simple life form level, and what’s good for one single cellular life form might benefit another life form, but the entirety of that first life form isn’t necessary for the second, so evolution would suggest that the absorbing life form will slowly whittle down what isn’t necessary.
Or has mitochondria always been perfectly fit for it’s function in our cells? (Hint it hasn’t)
I don’t think that Venter is suggesting intelligent design. He’s claiming, as a result of his research, that it’s not effective to assume simple explanations for genomics and especially for cellular biology.
Every technological improvement in the methods of research has revealed more complexity in organisms and so it behooves us to suspend dogmatic approaches to the genome. That’s the subject of the book discussed in the article.
Craig Venter is very controversial and his statements are provocative. I’m not qualified to critique the science in this field. But I’d recommend you to take a look at the work his team is doing with synthetic chromosomes and engineered cells.
Not an expert but it’s easy to see that information is not function. Like in computers, a sequence of bytes in memory can encode both operations and data. A single byte can be both. The two also mix up in dna, and adding a new random chunk of data to a mechanism like that will alter the expression, the fInal output. If an action must be repeated on all the elements of a list, and you add three random elements to the list, the result of the program changes. So no, it’s perfectly believable that there is no junk dna.
I’m sorry, but this is not computing, if it were you could think of DNA as an old spinning hard drive, sometimes you need to put pieces of data that will end up creating the program you’re going to run on different sides of the disc, fragmented memory if you will, you don’t need to read everything in a row to make the file, you need 8mb chunks there and there, there are start and stop codons that tell the RNA transcription proteins when to read and when to stop reading, and there are sometimes entire other genes between two sections of DNA that will eventually be “working in the same program”. There’s no need to read an entire strand of DNA, it’s not even done that way when the cells divide, it’s actually not possible, except in gamete production, to read the entire strand, because there’s a bit of extra (junk, telomeres) that cannot be read and reproduced, your DNA gets shorter every time your cells divide.
Structural similarities are most important (though still negligibly so) in recombination during meiosis, but even then the recombination is happening between strands of DNA of inherently equal lengths.
I believe you’re confusing DNA with protein formation when you’re saying the structure is important, there are many areas of DNA that have unnecessary lengths of extra codons. If you don’t believe this please look at plant genomes, there are some that are thousands of times larger in terms of base pairs, that have hundred times fewer genres.
See here for an overview of the current thinking on junk DNA: https://www.news-medical.net/life-sciences/Functions-of-Junk-DNA.aspx
Basically there are a variety of indicators that suggest A - Many “junk” regions are in some way evolutionarily important because many sections of “junk” are preserved across time, which, given the way DNA works, would almost certainly not be the case if they had no function, B - there is evidence non-coding regions actually do influence when and how other genes are expressed.
In essence, you could look at it as possibly metadata or maybe something very loosely along the lines of a Makefile for genetic code.
I think the assertion that it’s not “junk” is more alluding to the fact that although it does not directly encode proteins, asserting that non-coding proteins are “junk” requires us to ignore that there is clear evidence that non-coding regions likely do serve other purposes, or rather, requires us to proactively assert they serve no purpose simply because they don’t serve the very first purpose we hypothesized they should serve.
If a gene becomes disabled (a start triplicate pair gets changed to a nonsense triplicate), and it turns out that gene was no longer useful so there’s no impact on survivability/reproduction, what happens to the rest of the pairs before the next start triplicate? That stop triplicate and everything before it is now useless. Except evolution doesn’t understand useless, there’s just as much chance of flipping that gene back on as there is of shortening all of that non readable DNA by just 1 codon length, DNA replicators are very good at not dropping codons. But not you have a gene that isn’t being read (outside of replication) or transcribed, and it really isn’t costing the individual any significant amount of extra resources to continue to produce that set of code in that strand, so it just hangs out.
There are dozens of other mechanisms to control the rate of protein synthesis, why would junk DNA be the controlling mechanism for it when there are epigenetics, gates, chemical limits, so many different ways rare limit down the path.
“It’s there so it must have function” is spitting in the face of the theory of evolution. “It’s still in the genetic code so it must’ve been selected for” is barely less offensive. Evolution does not select for efficiency, it’s descent with modification, there is no pressure that says the genetic information must be as efficiently contained as possible. Example: https://en.m.wikipedia.org/wiki/Paris_japonica
Also I’m not at all arguing that proteins are junk (also not saying they’re peak efficiency, but “junk” in a ‘read’ section of DNA is clearly not ‘junk’), I’m arguing there are sections of DNA, especially repeating sections outside of start stop sections, that are without purpose.
This is a funny comment though, because “junk” DNA is involved with epigenetic regulation and cellular behavior.
“It’s there so it must have function”, “it’s still in the genetic code so it must have been selected for” is the least nuanced take,
“It’s there just randomly and therefore is junk”, and “evolution does not select for efficiency” is an improvement,
But “it’s there and it’s doing something despite not having a bespoke, prescribed function” and “evolution is a cascade of emergent effects and random chance, none of our genome is non-functional even though it is random” is the most up to date take
You seem like a biologist, why not go read some of these papers? Like the one I linked by Dan Nichols? Most people don’t have the background necessary to understand the language (no shade) but you seem to!
You can second-guess the current state of the literature on junk dna all you want but unless you have a research budget it’s a bit meaningless.
Also, I never said, and no scientist said, it was THE SINGLE AND ONLY mechanism influencing gene expression. Nobody also ever said the fact the “junk” is preserved was the CONCLUSIVE AND DEFINITIVE PROOF that it does something. Just that it’s a good indicator.
You’re not really responding to what the current hypotheses are.
Thank you for your answer, I will look up those things. Kind of an aside but regarding the dna getting shorter my undertanding was that it only happens when you get older and you don’t produce enough telomerase anymore that usually compensates the damage by extending the telomeres so the actual dna is not reduced during duplication.
Which theory exactly are we rejecting dogmatic approaches to?
There are several. One is the gene-centric theory of biology, which carries less weight in biology itself than it does in how biological sciences are communicated to laypersons - eg the Selfish Gene, which I could rip on for pages - and others include ideas that are considered contentious within biology, such as multilevel selection theory that extends beyond kin selection. I can’t begin to tell you about the number of arguments I’ve gotten into on that subject alone. I will frequently bring up that there is confusion as to what a “gene” actually is, and how it’s really determined by the context in which we’re using the word. There’s really just so much that needs to be re-evaluated.
" the Selfish Gene, which I could rip on for pages "
Please put me on the list to read that!
Ditto!
i believe the article suggests that the current way of communicating biology - that genes are the code that runs the machinery of life - is dogmatically adhered to by science communicators
it also suggests that when we communicate our new understandings that we are careful not to fall into another dogmatic theory, because it’s complex and we just don’t know
this is language used in the article, i don’t have enough information or understanding to know whether it’s true or not
Really good science news communicators (including many teachers because how are admins to judge?) are too rare … on YTube there’s … a half-dozen, maybe, at best?
Fraser Cain for astronomy
I linked to this in another comment, but figured you’d dig it.
https://www.subanima.org/veritasium/Sure large proteins have multiple forms and functions, but we do know the exact makeup shape and function of a lot of the smaller ones, like the ones that are small enough to get between either strand of an opened helix to read, transcribe, and send for translation the code for the larger proteins (anything inside the nucleus). Saying all proteins are inherently not understandable because there are large proteins that we don’t have full understanding of, is just as incorrect as saying we know every function of every synthesized protein.
Also I can’t stand veritasium, he gives over simplified, easy to latch onto, “explanations” of questions and misses out on a lot of minutia that would actually lead to a better understanding of why these questions exist. But people on the Internet love that stuff because they get an answer and are enabled to say “well actually” next time the topic comes up even though they’ve barely gotten any of the details of the problem itself. Pop science at it’s worst. Him and numberphile both
veritasium has also had some controversy around misleading half truths for click bait right?
i think it’s worth knowing the creator so you can appreciate it for what it is: not a source of truth, but a source of entertainment that if you find something fascinating you can look for other sources on… he does do entertaining videos with a science twist quite well, and uncovers some interesting topics… just not the be all and end all
There are way more than half a dozen, they’re just too dry to get the algorithm’s/communities attention. Most Internet viewers want an answer in x minutes or less, they don’t care to understand what the details of the questions are. If an answer is able to be given in a few minutes, why would the question exist for centuries?
It’s published in Nature or Science. Which means it’s better than the thruth (which we don’t have access to!), it’s high quality science.
Another metaphor that Ball criticizes is that of a protein with a fixed shape binding to its target being similar to how a key fits into a lock. Many proteins, he points out, have disordered domains — sections whose shape is not fixed, but changes constantly.
I dunno, kinda sounds to me like a good educational metaphor. Yea, not 100% accurate but good enough for high school biology. You need to make some simplifications for the sake of education. Not everyone can care about the complex intricacies of genes and proteins.
Good enough for high school biology. But not when you’re doing influential cancer research. The following is from Subanima’s article on the same subject:
One of the most influential papers in cancer biology published in 2000 was the “Hallmarks of cancer” by Douglas Hanahan and Robert Weinberg. It outlined six of the main capabilities of cancer and laid out a rough program for studying the disease ointo the 21st century. To date, it has over 39,000 citations which, in academia, is officially known as a shitton.
It was so successful that they released a sequel in 2011 which has over 62,000 citations - also known as a metric shitton.
But at the heart of both papers is the machine metaphor and the idea that if we just map out all the functions of proteins in one ginormous map, we’ll just have to run some maths and we’ll know everything we need to know to cure cancer. In 2000 they wrote:
Two decades from now, having fully charted the wiring diagrams of every cellular signalling pathway, it will be possible to lay out the complete ‘integrated circuit of the cell.’
He also notes the same thing you noted, that it’s a good metaphor for high schoolers.
I still feel like he’s nitpicking tbh, wiring diagrams can have devices with variable or probabilistic states and though the maths is very complex it’s theoretically possible to similate and map.
This maybe true, but these states aren’t being represented in the biological diagrams.
Why can’t we have both?
Edit: switched what to why.
I think we will. It’s still a useful analogy for initial understanding. However, I think we should be clear that it’s not quite perfect. Just like we have to be careful about bringing a Newtonian understanding into quantum physics where someone might believe a photon has mass because it has momentum.
deleted by creator
