Hacker News new | comments | show | ask | jobs | submit login
How Biologists Are Creating Life-Like Cells from Scratch (nature.com)
116 points by gotocake 4 days ago | hide | past | web | favorite | 39 comments

To clarify the exciting headline:

They’re creating cell sized bubbles out of cell like materials, and they’re able to get things inside them.

The opening paragraphs make it sound like they cells can split themselves, but what they’ve actually done is make a chip the can split them. Likewise construction of a cell is done entirely by chip fluidics to create what are afaict lipid bubbles, and they can use similar tech to force non spherical shapes.

You may need very simple mechanics / chemistry to achieve spontaneous division: https://en.wikipedia.org/wiki/Coacervate#History

One small but exciting step in the right direction.

Can you imagine in the future, we could have translations from logical code to biochemical pathways. You could specify the dependencies of a cell you are building in a similar fashion to (maven, npm, make).

Using bioinformatics data to configure your cells behaviour.

There would be compilation errors were cells would not be stable or materialise, and runtime errors in which a metabolic or other type of pathway fails, causing the cell to throw an exception (dies/ or emits other action).

I would truly relish the moment when my cell gets a virus.

Seems like synthetic biology is a field you should look into! It's very exciting!

Ill need to look into it.

My background is in Biomed but I joined the dark side and became a computer scientist instead.

Slightly OT but related interesting trivia. Look up a language called CRN++ developed by UT Austin for "programming deterministic (mass-action) chemical kinetics in performing computations". Press release type article here - https://interestingengineering.com/synthetic-biology-gets-a-...

ArXiv abstract here: https://arxiv.org/abs/1809.07430

Git repo here: https://github.com/marko-vasic/crnPlusPlus

Hope to see some interesting synthetic molecules published by you :-)

You may also be interested in the field of biological computing (https://en.wikipedia.org/wiki/Biological_computing#Biochemic...). Since your background covers both biology and computer science. Plus the field is new enough that you could be a part of some exciting research.

I wouldn't call it the dark side, I think there can always be an overlap in one's skillset.

having been a biochemist (and maybe going back) but also a professional programmer, I am very pessimistic about this.

One time, though, I used ruby to scrape a chinese website where they made 'public' the genome of an organism. They clearly weren't interested in actually making it public (as the sequences were very hard to pull), and after we had gotten the data they made it impossible and took down announcements that it was public, so sometimes real computer skills are welcome!

The article speculates that it will be hard to get these synthetic cellular systems to evolve, even if they manage to self replicate and survive given their initial rational design. I don't see this as a problem if they really can be made self sustaining. If the cells replicate then they are already implementing an evolutionary process. We just need to be sure that mutations can arise due to replication error or DNA damage. But if they don't replicate then human designers mediate their evolution. This just seems so straightforward to me. Am I missing something?

If they don't replicate, they aren't alive. Replication is the defining characteristic of life.

If we're going to split hairs like that then blood cells and the neurons of some species of adult fairy wasps are not alive.

And if you response is that the rest of the organism fulfills the self replication process through sexual reproduction, then why can't these cells being assisted by an outside process like the splitting chip be considered alive in whole like an animal is?

> blood cells and the neurons of some species of adult fairy wasps are not alive

That's right, they aren't. You hair and fingernails aren't alive either. Nor is your clothing despite the fact that that too is part of the human phenotype. [1]

> why can't these cells being assisted by an outside process like the splitting chip be considered alive in whole like an animal is?

Because the splitting chip isn't produced by the cells, it's produced by humans. Take the humans away and the whole process comes to a grinding halt.

[1] https://en.wikipedia.org/wiki/The_Extended_Phenotype

So build a lipid membrane, insert some DNA from a bacteria in there, and it will just function? Sounds like they missed explaining how the membrane will extract and execute the instructions from the DNA... which in itself is a huge question mark.

They mentioned they included some cellular machinery involved in transcription and translation, and were able to show that "cells" produced fluorescent protein in response to an environmental trigger

"Bacteria have no cytophysiology, only chemistry", as someone said.

But of course you'll need some transcription enzymes to make RNA and then proteins, and plenty of raw amino acids floating around.

Even a single cell is more complicated than anything humans have ever created. Clickbait title?

I wonder. A quick google suggests that a typical cell might contain 10^14 atoms, and that the number of transistors in a current supercomputer is of a similar order of magnitude. Is the cell really "more complicated" than the supercomputer? The nature of the complexity is so different that I'm not sure the comparison is all that meaningful.

> Is the cell really "more complicated" than the supercomputer?

Let us ask a different question that may suggest a possible answer.

How many different dynamic manufacturing processes occur inside a supercomputer as compared to the number of different manufacturing, transportation and communication processes that control those manufacturing and transportation processes occurring within any single living biological cell?

Cells have a rather limited set of manufacturing processes.

Mycoplasma mycoides only has 525 genes and we can cut tha down to 473 and end up with something that self replicates. https://www.nationalgeographic.com/science/phenomena/2016/04.... (I find it amusing how breathless they talk about the fact their where some unknowns in that list rather amusing.)

Let me ask some more questions and see what insights we can obtain from them.

Within each manufacturing process, how many different kinds of steps are there and how "complicated" is each step (the required processing that is required to do that step)?

What kind of dynamic manufacturing infrastructures are created and then taken down within the cell for each of these processes?

That’s an ambiguous question. If you want to accurately simulate matter, the larger the block of matter the more processing power it takes. We think of DNA as more complex than a crystalline structure becase crystals have more order, but that’s abstracting away a lot of the details. Which means manufacturing a super computer is vastly more complex in terms of matter.

In some ways we better understand how super computers are made, but we can also far more easily replicate cells than a supercomputer. So, again supper computers are more complex.

In terms of the minimum amount of information to turn raw matter into a simple cell, again super computers take more information.

I suspect you want some sort of third definion where cells are more complex, but that’s more begging the question than how things actually are.

When you get the chance have a listen to some of the discussions about organic chemical processes occurring in cells by Dr James Tour. From an engineering and manufacturing point of view, the processes that occur in a cell are far beyond what we are able to achieve in any of our engineering and manufacturing processes.

It is not about simulating processes, it is about doing those processes.

We can manufacture DNA from 'scratch'. That means as long as we have the sequence we can largely use those same processes. Sure, we don't necessarily understand what's going on, but neither did our ancestors you first manufactured steel etc etc.

Protean folding is very complex to simulate. But, so is even a hydrogen atom starting from quantum mechanic equations. Saying we can abstract away that hydrogen atom does not mean the complexity disappears. Further from a QM standpoint protean folding is really slow, it's a complex dance like plate tectonics even if it's blindingly fast from our viewpoint.

By nature, we understand the "what' and "why' of each supercomputer component, as they have been built by us from the bottom up. We are orders of magnitude from understanding each component of the cell. When you get into the weeds, the methods used to dissect these processes are generally slow, indirectly and subject to relatively subjective interpretation and poor reproducablility (vs chemistry or physics). Don't get me wrong, I love this stuff and was drawn to a molecular biology career by curiosity around the wide grey line that separates what is living and what is a bag of molecules, but having spend many years at the bench still think that the Biology part of Synthetic Biology will be the bottleneck rather than the bioinformatics/synthetic methods, tho improvements in this stuff will greatly help.

Exactly. Look at the computational power we have thrown into figuring out just how a protein folds for example. I worked in a lab where we used a lot of inference in the 2D structure of RNA over a decade ago after the 23s and 16s ribosomes tertiary structures were resolved. I was simply working with 70-120 base-pair tRNAs and that alone was computationally complex using a known model to compare against! We have still but scratched the surface.

Building a core cell always intrigued me as an undergraduate and I was often interested in the bare minimum (DNA-wise) lab manufactured bacterial strains that were used in experiments.

How many types of atoms are there vs types of transistors? How many combinations and permutations can the atoms rearrange themselves into vs transistors? How many other atoms can an atom interact with in a given time period vs transistors? How dynamic is the state of a cell vs a supercomputer, and how diverse are the state changes?

I don't know the answers to any of these but I think that a comparison of complexity should take more than one metric into account

I'm working (as a consultant) for a company that does state-of-the-art chip design. The complexity is mind-boggling. The number of different types of core components isn't very high; at root it's all transistors and (nominally) rectangular pieces of metal. But these simple building blocks interact with each other in ridiculously complicated ways. Just processing the files that contain the timing information for the current generation of chip fabrication technology takes over an hour on a fully-tricked-out server. Checking a design to see if it meets a timing spec takes many hours using dozens of servers. And then you have to deal with power, thermal constraints, geometric design rule checks (the list of rules you have to follow is a PDF document hundreds of pages long), clock distribution... Frankly, it amazes me that state-of-the-art chips work at all.

How are you drawing the comparison? How many complexity-units would you rate, say, the economy of Morocco?

As humans, we don't design economies. Instead we generally try to steer them and fail. Ref: Governments printing money and no gold standard.

To give their life worth, some arrogant economists/finance people think they do have control. But they don't.

> no gold standard

What make you think that a gold standard would be intrinsic to the value of money?

Nowadays the value of money is represented by peoples debt, i.e. work that will be done in the future.

Printing money just lets the government increase inflation. It doesn't destroy value, it just "steals" from the people, espacially those with much savings.

That's the point though. Without something actually tangibly valuable (IE gold), backing every paper dollar, there is no real value to the dollar anymore. There is value, but its constructed value. And we all know what goes up must come down.

Don't forget, the dollar in your pocket is just a piece of paper.

EDIT: Yes, it is all a con job. One day we will realise.

You are begging the question of gold (or anything, really) being "tangibly valuable" versus having "constructed value."

For something to have true value, it must be limited in supply and also people must have a common agreement that it is valuable.

Money used to be like this before the Gold Standard was removed. When that happened, money lost its true value. Off course, money people could get creative then.

Now the whole system relies on the population simply "believing" that fictional and unlimited entities (dollars) are valuable.

Side Note: The reason why humanity has chosen gold to obsess over as an ultimate form of money, is not relevant to the fact that it has been, is, and always will be. See: https://www.marketwatch.com/story/why-china-and-russia-are-b..., for one of numerous reports of governments buying up as much gold as they can.

Was Morocco "created". I would think its fairer to say that it "evolved".

Though its an interesting question.

> And as the engineers of synthetic life, she and her colleagues can easily incorporate controls or a kill switch that renders the cells harmless.

That's a bold claim given how evolution routes around failures.

yeah. that line made me laugh. it's the kind of thing i expect to see in a low-grade sci-fi horror screenplay.

Mary: "But to make all these pseudo-cells reproduce at room temperature, we had to mimic properties of the polio, smallpox, and ebola viruses, professor! What if it gets out of control?!"

Professor Devin Daylooded: "Don't worry, Mary Goodsave. You may be a super-intelligent grad student who sees things I don't, but I will ignore your cautionary message which even the audience understands because I can easily incorporate controls or a kill switch that renders the cells harmless."

This work focuses on the membrane, but I'm more concerned about the feasability to synthesize ribosomes from scratch. Ribosomes are incredibly complex tiny machines.

Reminds me of a simulation of the formation of a lipid bilayer:


(the molecules that suddenly appear/disappear are just moving in/out of the plane of view; a camera focus/blur effect would have been nicer, I think)

Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | Legal | Apply to YC | Contact