The second decade of synthetic biology: 2010–2020 | Hacker News Reader

The second decade of synthetic biology: 2010–2020

nature.com

108 pointsIce_cream_suit19 comments

I've been in synthetic biology / DIYbio for the last ~8 years, and recently worked 3 years in Drew Endy's lab and got a decent perspective on how things went down throughout the years.

If there is something the computer hackers should know about synthetic biology, it's that it is in its infancy (ie, there are pretty few players who are leading all those projects listed). It is really early. Joining open source software projects to help synthetic biologists will non-trivially affect the future of the field (they don't know how to program)

If I had to predict what is coming in the next 10 years, I'd say that improvements in biotech networks are going to enable cool stuff. The BioBricks Foundation's Bionet Project (worked on that a little) was a good foray into this idea. Basically, networks of materials and software are going to cause the price to participate to start dropping into early home-computer hobbyist realm.

Synthetic biology right now is in very infant stages of networking. We have Addgene, which is alright, but large scale collaboration on important projects are really unheard of. I think that'll change.

I don't think we've had the combo of visionary (Drew Endy, Tom Knight, etc) and pragmatist in academic synthetic biology yet to move forward on important projects. I think a new a group of synthetic biologists are incubating right now that may change that.

> Joining open source software projects to help synthetic biologists will non-trivially affect the future of the field (they don't know how to program)

Can you share some projects?

Sure! One thing I have to note though - the field is fundamentally different than software engineering. In software, it isn't unusual for someone to maintain a software package for a long time, and for other people to contribute. In biology, typically software projects are made by individuals (only supported through their graduate school) and are never contributed to by outsiders (most don't know how Git development and PRs work). There are lots of projects in the normal bio category, but probably aren't useful.

The 2 that come to mind that are a bit more organized and in development are SBOL (https://sbolstandard.org/) and Poly (https://github.com/TimothyStiles/poly, self plug).

SBOL is a project looking to make a better format for sharing DNA sequences. For context, the same format (GenBank format) has been used as the standard for sharing biological sequences since 1982. As you can imagine, our understanding of DNA sequences has improved since that point, but we have no way to encode much of that information in a shareable format. So it doesn't get shared. That means a lot of genetic elements can't be computed (data is there, it isn't shared) so synthetic biologists just keep it floating in their heads to make experiments work. SBOL is a new format that has been running for about 10 years trying to change that - it is pretty much the definition of "design by committee", but their design is extremely thought out, and they have the community reach necessary to get acceptance. However, the tools just ain't there yet to use SBOL for widespread adoption. I do believe it'll get there though, but they don't have enough skilled programmers.

Poly, which is a project I've been involved, is aiming to make good CLI tools for biotech (in addition to a kickass Golang library). It's fairly early, but ran by an actual software engineer, and we're making pretty good progress on tooling, and a decent community of biologists. For example of some stuff we're working on right now for version 1.0.0 - universal identifiers for genetic sequences, JSON representations of GenBank files (for interop), codon optimization / DNA synthesis optimization, primer design, and plasmid cloning simulation. One of the exciting parts of the project is an algorithm I designed that is able to extremely efficiently index genetic sequences for search. The goal is to index ALL public genetic sequences and make them searchable through a simple API (Should get a 100-1000x speed up to what is currently available).

SBOL does calls every Wednesday with updates to their work, and Jacob Beal kinda runs the show (awesome guy). Poly has a gaming night every Friday, which is a good way to figure out if the community is right. We're trying to get more issues on Github, but that is a work in progress!

I would like to know which open source projects are used by synthetic biologists.

They aren't many (most of it is closed source). The one package that is used a stupid amount is the Biopython package. Most synthetic biologists I know are pretty much only trained in python (because bioinformatics), so use that package religiously.

As examples of closed source software that there aren't good open source alternatives to, see Benchling and SnapGene. No good plasmid editors are open source and hackable.

Joining the call for projects to help with. Also, what bio communities are best to inquire with about help they may need?

^ Listed the 2 that I think really need help above

Biology isn't nearly as organized as software, so the bio communities are a bit disparate. It also doesn't help that biology has such a large scope. From my experience, conferences and academic labs seem to be the only real places were bio folks congregate. To get a sense of needs, conferences are definitely the way to go. iGEM comes to mind as probably the best one to attend.

The culture is a little bit different, too, which might impact that. The idea of someone coming and saying "hey I have some <engineering> skill, how can I help?" is pretty foreign. Contributions are almost always from people within a lab, so there aren't any mechanisms in place to get interested outsiders involved.

I think that'll change with the increased networking in biology, but right now, people just aren't helping each other very much :(

I'm very excited to see where we'll be in another 10 years. The possibilities in terms of medical and industrial applications seems hard to overstate.

Related, but one of the best talks I've ever seen was this 2018 talk on the subject. It gets into how organisms seem to have bioelectrical patterns that store information separately from genomic, anatomical states. These bioelectrical patterns play a huge role in developmental processes, as they appear to control how the organism is structured at a high level. As a layman and a programmer, to me it sounds like the genome is the actual implementation of a function and the bioelectrical pattern is the sequence of functions you want to execute. They're able to manipulate these bioelectrical patterns to do things like replace a worms tail with another head. It's absolutely crazy to think about what could be possible here.

https://www.youtube.com/watch?v=RjD1aLm4Thg&t=0s&index=6&lis...

It's worth taking a look at some of the simulation software supporting this field. Lots of potential.

"Bioelectric Tissue Simulation Engine (BETSE). Bio-realistic modelling of dynamic electrochemical phenomena in gap junction-networked cell collectives with a focus on spatio-temporal pattern formation."

https://gitlab.com/betse

It's pretty hard to compare biology to programming, mainly because analogies start to break down and it becomes fundamentally different. The genome is just a collection of atoms that has figured out how to trick other atoms do work (in order to make more copies of itself). And it will do weird, weird things to accomplish that task.

Forgive my limited knowledge of biochemistry but isn’t the gist of it that nucleotides code for and combine to form amino acids which when put in certain sequences form enzymes and proteins whose structure then enables some function so it is much like the paper tape model of a computer except instead of paper with holes in it you have a polymer with a few types of “hooks” that allows it to fold into an origami machine that can manipulate or join with other bits of paper to do things? This is different from programming a computer which relies on reading instructions from such a tape but only in that the computer and its functions/actuators are made of tape. It may be a strained metaphor but it is how I conceptualize it. Please correct me if I’m wrong.

That's is essentially correct. But even if we discount the steroids this "origami machine" is on, and assume is an un-fancy origami machine, it still defies easy reasoning. Or at least, the kind of reasoning we are used to when programming digital computers.

Absolutely, I don’t deny the complexity! I still wonder at the mind boggling accomplishments of evolution. It’s like being able to create a working car out of nothing but wet spaghetti!

Discussed at the time:

https://news.ycombinator.com/item?id=18700328

https://news.ycombinator.com/item?id=18736698

Wow, thanks for this. Around 2005 or so I was super into synthetic biology but have paid almost no attention since. If you told me this was a speculative "letter from 2020" written back then I would almost believe you, it's so jam-packed with the kind of thing we were expecting. (The obvious stuff, anyway--that the Baker lab's protein design work would be an enabling technology. That some kind of design parallelization was needed to replace hand-construction of gene circuits. That logic and sensors would find practical applications in something like cell-based therapeutics.)

Interesting thing about these revolutions and their makers. Back then I actually went and did a first year grad school rotation in what I thought of as a synthetic biology lab--they had published one of the two back-to-back Nature papers that kicked off the hype in 2000. Surely this was the place to be! But pretty quickly I was confused and puzzled by the lack of any late night bull sessions about the future of synthetic biology, the lack of interest in the stuff that Drew Endy and Tom Knight were doing, and the lack of anyone mentoring newbs in the secrets of the field. People who had done synthetic biology projects seemed to be wrapping up or leaving. (Though I later learned that one guy who almost never talked and worked by himself late at night was doing kickass work--he eventually became an MIT professor.)

I even saw the lab attract another wide-eyed synthetic biology enthusiast--who coincidentally I knew from del.icio.us ( https://news.ycombinator.com/item?id=23986789 )--as a junior year intern, assign him a non-synthetic biology project, and then regard him as kind of a flake for being so interested in synthetic bio (he later founded a science related nonprofit that got some press.)

Anyway, it took a while for it to really sink in that the same lab could help kick off the field and yet not be "all in". Turns out the lab was basically a vehicle for its ambitious and very politically savvy PI to branch out into different fields. Which is a fine environment to work in! ...For other, more career-minded people, especially ones who bring their own expertise to the table (as postdocs, etc).

Forthcoming paper from Church Lab gives a peak at a possible future. First line: "Xenotransplantation, specifically the use of porcine organs for human transplantation, has long been sought after as an alternative for patients suffering from organ failure"

Extensive Mammalian Germline Genome Engineering

https://www.biorxiv.org/content/10.1101/2019.12.17.876862v1

Can a synthetic biologist comment on whether or not something like mitochondrial base editing should have been included. If it should have, is its exclusion presumably due to the fact that it was published so recently?

I think this is a nice article, but I can't help but be frustrated by the stochasticity of HN since I posted it yesterday :D

Synthetic biologist here - I don't think mitochondrial base editing should be included. The basic technology that enabled it, TALENs, were actually mentioned. Mitochondria is just a specific use case of that very powerful technology.

Hm, I thought this paper out of the Broad came up with a novel method. I haven't actually read it, but I saw a lot of press about it at the time.

https://www.nature.com/articles/s41586-020-2477-4

Reply on news.ycombinator.com