Hacker News Reader

Comment by grishka | Hacker News Reader

Whenever morphogenesis comes up, especially with the usual "work from DNA and proteins up" approach, I feel obliged to mention Michael Levin. He also researches morphogenesis, except his approach is on a higher abstraction level. He manipulates the electrical network that cells use to communicate with each other to organize their work towards multicellular goals. Do watch his lectures, it's some really fascinating stuff. You don't need to touch genes at all to regrow limbs and organs. Neither do you need to operate on that low level to cure cancer. This whole "let's effect organism-level changes by observing and manipulating DNA and proteins" feels to me like "let's add a feature to a react web app by observing and modifying the behavior of individual transistors in the CPU". Doable given enough time, but you'd drown in complexity in the process.

https://www.youtube.com/watch?v=Ed3ioGO7g10

View on HN

I think that analogy is very apt- the fact that we have such a hard time understanding and predicting the outcomes of bioengineering/synthetic biology suggests that we are still missing huge parts of the picture of how life works.

It seems very likely to me that our views are too reductionist, and the much of the key information isn't even encoded at the level of DNA as previously assumed.

The cells multicellular organisms are constructed from are all shockingly similar... "cells" are basically a solved problem for this type of life, and somewhat frozen in their functionality because major changes would disrupt the larger systems that they make up.

> The cells multicellular organisms are constructed from are all shockingly similar.

They are not. They vary greatly in structure and function.

What I said was a matter of opinion, I didn't quantify it in any way, so I'm not sure how you can say that, but I will elaborate more so you have something to really disagree with :-)

Sure, of course the different multicellular eukaryotes vary, but the diversity is very low compared to what we see across all life, look at Figure 1 here for example [1].

I am looking at it from the perspective of the structure of the metabolic network, as this is my area of research. There is a huge amount of diversity in the fundamental structures of metabolic networks, types of metabolites and enzymes produced, etc. across all life, but from that perspective the differences between one multicellular eukaryote and another are slight in comparison. Multicellular eukaryotes have evolved very different behaviors, shapes, sizes, etc. from a relatively conserved biochemistry. When taken out of the context of an organism in cell culture, the cells are similar enough that their behavior under diverse experimental conditions can be predicted by nearly identical metabolic models, whereas you would see nothing of the sort with distantly related bacteria- the metabolism can be massively rearranged.

[1] https://www.nature.com/articles/nmicrobiol201648

You made a broad, almost dismissive generalization from, as it turns out now, a purely macroscopic view. It's like a physicist studying galaxy filaments saying "all galaxies are shockingly similar", because the filaments show more complex structures than individual galaxies.

Which is exactly the point- a systems view offers a different perspective than the reductionist view.

Piling on; I emailed Michael Levin during lockdown and he replied with answers to some high level questions and a bit of documentation. Not sure what his pace is like these days but he seems open to sincere attempts at discourse.

Let's fund this instead of shitcoins, layers of indirection around POSIX.

I can't see any way that you could regrow a limb without "touching genes". And Levin doesn't say that in his video. At the very least, limb regrowth would require synthesis of many mRNA and their protein products, under the control of transcription factors.

Yes, but you don't do that. The cells themselves do it. You only trigger that mechanism and you do it by opening and closing specific ion channels in specific patterns.

The article is specifically about a difference in the mTOR protein, so in the end, a genetic difference. You will not trigger a limb regrowth in humans by opening and closing specific ion channels in our cells.

> You will not trigger a limb regrowth in humans by opening and closing specific ion channels in our cells.

Levin's research begs to differ. He did successfully regrow an amputated leg in a frog.

I seem to think you're interpreting that in a way that is not consistent with our current understanding of limb regrowth.

This is the paper: https://www.science.org/doi/10.1126/sciadv.abj2164 and it clearly states they treated a frog (xenopus, a common model organism), which does not normally regenerate when it's an adult. The paper clearly states that they treated with a set of factors that they believe induced a developmental plan- implemented by the transcription of DNA, translation of RNA, and expression of proteins. The initial treatment was removed, and the effect continued.

Also there's nothing about ion channels in this- the factors they used must have been actively transported- they bind to protein receptors.

I’ve watched at least half a dozen long presentations and interviews from dr levin and I’d say the previous commenter is essentially repeating it exactly how Levin describes it. To be fair it’s not exact. What I understood from the interviews is that they have been imaging bioelectrical activity in organisms during development and claim to have deciphered patterns associated with growth of certain structures. He talks about a certain ion channel drug (one used for epilepsy I believe) which they used to reproduce the bioelectric field previously observed and the organism grew a leg. Multiple legs actually, wherever they wanted. Also working eyes on arbitrary locations.

So from a high level they say they have figured out how to “speak” to cells in a sense and instruct them to start building certain structures. And they emphasize that they aren’t making any changes to the genome or manage any of the low level processes themselves.

Unless you can point me to some succinct papers or summaries, I'm going to continue with my believe that the dog wags the tail, not the other way around.

Frogs are not humans. Amphibians do have the genetic potential for limb regrowth.

Genes encode proteins and do not contain any other information.

The celular machinery and its goals are separate from genes and can be controlled without "touching genes".

"Genes encode proteins and do not contain any other information": this is entirely false. It's well established that genes contain non-encoding regulatory regions (which happen to be very important for the process we're discusion).

All the cellular machinery ultimately derives from genetic regulation and protein expression. You can't make more of it wihtout making more mRNA, translating that to protein.

This is bio 101, basic stuff, unless I'm completely misreading what you're saying.

The code is there. It's a timing and signalling problem.

What? Michael Levin? How about Dr. Robert O. Becker? I read about something similar in the 80s. Here's a link to one of his books, The Body Electric: https://a.co/d/fqzpaIv

Does Michael credit Becker for any of his work?

Googled. Yes he does! You've made my day! So happy to see someone continuing Becker's work.

Reply on news.ycombinator.com