Glowing Plant Is One of Y Combinator’s First Biotech Startups | Hacker News Reader

Glowing Plant Is One of Y Combinator’s First Biotech Startups

techcrunch.com

100 pointskatm61 comments

I'm one of the founders of this project, and will be available to answer any questions you have.

What general techniques are you using? Perhaps another way to put it: which Wikipedia article is most related to your methodology?

We use a broad range of technologies: 1. We start with protein engineering to improve the function of the enzymes which make the glowing reaction. We are using a process called directed evolution to achieve this: http://en.wikipedia.org/wiki/Directed_evolution 2. Then we recode the genes coding for our enzymes so that they can be read by plants - here we test a large number of combinations. This uses a whole suite of technologies: Genome Compiler software to design the sequence, DNA synthesis of the parts, DNA assembly using transcriptics and then transient experiments to test the DNA sequence in plants as described here: http://en.wikipedia.org/wiki/Agroinfiltration 3. Finally when we have a DNA sequence which is working well we insert that into the plant germline (so that it's inheritable) using a gene gun:http://en.wikipedia.org/wiki/Gene_gun

How much control does the gene gun process give you over where the new sequence ends up? Is it in the same place in all of your production seeds or do they structurally differ? I guess a related question is if all the seeds come from one ancestor cell where the insertion was done or if you do the insertion multiple times?

(I work with human genetics and am fascinated by the prevalence of structural variation.)

The gene gun inserts randomly. We do a large number of insertions and will then test to see which are the best plants and take those forward to production. All the seeds shipped will however come from a single parent.

How bright is the glow?

At the moment it's comparable to glow-in-the-dark paint, imagine those stars many people had on their ceiling as kids. we are working on improving the luminosity however, we are improving it about 10-25% per week using a process known as directed evolution.

Aren't there some pretty fundamental physical limits here? This wikipedia article: http://en.wikipedia.org/wiki/Incandescent_light_bulb claims that the luminous efficiency of a low-wattage incandescent bulb is maybe in the neighborhood of 1%. So that suggests that to get as much light as a 25 watt incandescent bulb, assuming that your plant is perfectly converting input energy to light output, you'd need to be continuously burning 0.25 watts. Can a plant possibly metabolize or store that much energy for long?

Bioluminescense is known as cold-light because it's so efficient - basically zero energy is wasted as heat. There will be fundamental limits (currently unknown), most likely from toxicity or waste products hurting the cells, but we are far from pushing those boundaries yet.

Right, but a good first order estimate woudl be to take the brighest bioluminenscent molecule, pack it in a sphere with volume = your plant, and compute lux at 3 meters. Or something like that. Without growing lensing structures or other ways to focus/point light, the r3 falloff is gonna kill you.

Shouldn't the falloff be proportional to 1/r^2, instead of 1/r^3?

1/r^3 = 1/r^2 (falloff in vacuum) * 1/r (light absorption by air)

point sphere radiating a point at distance r. I'm not doing the calculation the way a physicist would, or rather, I'm calculating the amount of EMF falling on a point, not the energy of the full surface. The reason being, a person standing and looking at something cares how much illumination falls on the destination point, not the total illumination of the scene.

Ah, duh. Had to go back and look up some basic geometry.

So, surface of sphere is proportional to r2. At distance r, with 1/r absorption, the illumination of a spot on a sphere at distance r would fall off as 1/r3. But I wasn't including absorption, so my original claim of 1/r3 due to spherical emission was wrong.

Typical crop plants are ~1-2% efficient at gathering sunlight. You get ~100 watts of solar energy per square foot for ~8 hours a day. So that's 8-16 watt hours per day our .33 to .66 watts per hour so it's at least in the right ballpark.

Note: Sugarcane peaks at around 7-8% efficient so with the right plant .25w seems doable.

http://en.wikipedia.org/wiki/Photosynthetic_efficiency

Congrats! Send Kyle a high five for us, it's been awesome watching your progress!

What plants are you working with? Are they pet and kid safe? Even at your current brightness levels it could make a cool nightlight for kids, as long as it doesn't kill them or the cat.

We work with Arabidopsis. Arabidopsis is not very tasty but is not toxic either.

Do you have pictures of glowing plants? (just been able to see 2 lowres photos so far)

Here you go (note this is a 30 second exposure): http://cdn.optimizely.com/img/536610845/38f49c2ac92c415d824b...

Is that with or without luciferin added?

is your plant hybridized, or do you plan on developing hybrids to increase vigor after you've gotten the glow bright enough?

You said elsewhere that you're increasing brightness 10-25% per week, which seems crazy fast. How fast do your plants reach maturity?

Increased luminosity per week comes from protein engineering the pathways in bacteria. We then test those improvements in plant leaves, which takes about four weeks, before inserting the full DNA in a plant which is the slow part which takes 6 months.

Will you eventually ship seeds overseas?

We certainly hope so. Due to the divergence of regulations in different countries it will be on a country by country basis, most likely starting with Canada and Asia.

How do you guarantee that this DNA will not get released into the wild, and what is your responsibility for the consequences if it does?

Honestly I wouldn't be upset if the truth was none and none. May as well be happening in the open.

edit: If I got these I'd be breeding it with all sorts of other close species; see what sticks.

Not the op, but I have an answer to this.

This plant is wasting energy by glowing. It serves no purpose that enhances the plant's fitness outside of a human environment. If it were released into the wild, chances are it would lose out in natural selection vs wild competitors.

The same is true of most human domesticated species, most of which are effectively genetically engineered through ages of selective breeding. Picture dogs vs wolves, cows vs wild buffalo, etc.

Exactly, the plant is not being given any kind of selective advantage and the genes in fact incur a metabolic cost on the plant so it will lose out to natural selection. The domesticated analogy is a great one, plus remember we took the genes from the wild in the first place!

Exactly, (...)

Seriously? As the CEO of the company and the person ultimately responsible for this project, it is troubling that you endorse these simplistic and erroneous views of plant biosafety. Fitness is notoriously ecosystem-dependent and hard to predict, which is why microcosm and field experiments-based risk assessments exist in the first place --it unfortunately can't just be eyeballed from a simple metabolic account like that. And either way there is the additional risk of transgene flow, which is even more long-term and less understood --especially for such relatively distant horizontal transfers.

You might have (unaccountably) skirted APHIS regulation, but that doesn't mean you don't have an ethical obligation to (1) thoroughly assess the biosafety of these plants; and (2) be frank in communicating these risks (and their uncertainty) to the public, even if they do not make as clean of a narrative as one would like.

One of the things we try to do with this project is simplify things so that people can more clearly understand. Intuitively it's not clear why there are risk from our work (beyond the unknown unknowns). We have engaged in a number of panels with expert ecologists and after one an MIT professor asked why we were always talking about glowing plants as there were no serious risks here. You can read one of the reports here: http://www.wilsoncenter.org/sites/default/files/SYNBIO_res_a...

Also, specifically to your point about transgene flow. Higher level organisms (eg animals, plants) very very rarely exchange DNA - this is a good thing or else you might start photosynthesizing after eating salad for lunch. There are a couple of recent papers saying maybe (and that's maybe in a scientific probability, ie small probability) this happens on the order of millions of years. APHIS doesn't even look at this issue anymore, here's their comment on it:

Potential impacts from transferring genetic information from plant to organisms with which it cannot interbreed • First, many genomes (or parts thereof) have been sequenced from bacteria that are closely associated with plants including Agrobacterium and Rhizobium (Kaneko et al. 2000; Wood et al. 2001; Kaneko et al. 2002). There is no evidence that these organisms contain genes derived from plants. • Second, in cases where review of sequence data implied that horizontal gene transfer occurred, these events are believed to occur on an evolutionary time scale on the order of millions of years (Koonin et al. 2001; Brown 2003). • Third, transgene DNA promoters and coding sequences are optimized for plant expression, not prokaryotic bacterial expression. Thus even if horizontal gene transfer occurred, proteins corresponding to the transgenes are not likely to be produced. • Fourth, the FDA has evaluated horizontal gene transfer from the use of antibiotic resistance marker genes, and concluded that the likelihood of transfer of antibiotic resistance genes from plant genomes to microorganisms in the gastrointestinal tract of humans or animals, or in the environment, is remote (Council for Biotechnology Information, 2001; http://vm.cfsan.fda.gov/~dms/opa-armg.html, accessed 1/26/10). • Finally, a recent review of issues related to horizontal gene transfer concluded that this type of gene transfer is unlikely to occur and poses negligible risks to human health or the environment (Keese 2008).

Intuitively it's not clear why there are risks from our work.

Far from it! The very workshop minutes you are citing [1, Appendix 3] directly contradict your point, with participants raising a number of serious concerns about your project:

"Many questions focused on the ethical and philosophical issues associated with the project, including matters of consenting public and responsible science practice. However, only questions relating to the research agenda, like those impacting regulatory consideration and potential ecological effects, are included here.

Ecological interactions:

- How will the impact of the bioluminescence from the plants on wild organisms be tracked across seed destinations? If this can’t be tracked or known ahead of time, then how is the application ready for release?

- How have the interactions of insects with the glowing plant been characterized? Does this have the potential to disrupt pollinators?

- Arabadopsis is frequently used in laboratories specifically because it is easy to grow and is a weedy species. How does this align with comments made regarding the difficulty of growing the plants, and the extreme unlikelihood that the seeds would take root and grow if released outdoors?

- Were any types of biocontrol mechanisms employed or tested in the system? Why was sterility not introduced into the system when it could serve such a purpose?

Determining regulatory coverage:

- Habitats of relevance.

- Tests for characterizing the application: The seed packets will not be regulated because the seeds were created using gene guns, while the DIY maker kits will be regulated because the system will rely on Agrobacterium. Neither can be shipped internationally. However, once these are distributed, how will use be monitored? If the plants are expected to be sent to thousands of individual sites around the country, how are habitats of relevance being determined? Are all of these locations being tracked, characterized, and assessed for specific vulnerabilities in advance of product release? The USDA APHIS test framework is insufficient for this purpose. How could a model be developed to evaluate such a widely distributed application? What questions would need addressing in order to appropriately characterize the effort?

- Demonstration of impact: Compare the altered plants to other mustards, and assess how well they grow. A series of greenhouse competition assays in a variety of environ- ments would be a good start, and any identified differences could point to areas requiring further study. Also study known pathway interactions up- and downstream. Emphasize the study of the resulting phenotypes, not the genetic modifications."

[1] http://www.wilsoncenter.org/sites/default/files/SYNBIO_res_a...

In regards to your "response" to the risk of transgene flow, frankly I don't even know where to begin...

First, what you cite is not a comment from APHIS, but a verbatim quote from an earlier post from your company blog [1]; which is supposed to be a draft outline of proposed biosafety tests, but which largely consists of paraphrases and language from two already completed APHIS plant pest risk assessments for a completely unrelated rose variety using plasmid transgenes, some of whose conclusions and comments you then use to (wrongly) extrapolate about your product and its risks. This is highly misleading, to say the least.

Second, your statements about transgene flow risk are outdated and mischaracterize current knowledge. Suffice to say, your technical team (or even any of the experts you engage with, say, in SYNBIO 7) is more than capable of briefing you on the current re-assessment and state of the art.

Best of luck with your project, but do keep in mind ethical obligations to your stakeholders override any other imperatives or agendas you might have.

[1] http://blog.glowingplant.com/post/57840496382/tests-we-plan-...

> How do you guarantee that this DNA will not get released into the wild, and what is your responsibility for the consequences if it does?

Don't be silly. It's only in movies that monsterous mutations appear.

What consequences could this kind of useless, extra metabolical burden (luminescence) have? It would quickly disappear in nature.

I've wanted to do this since I saw a picture of a glowing tobacco plant in my biology textbook back in 1995.

I've since done quite a bit of research on it, and there are some pretty serious obstacles. I assume you want to use luciferase... are you going to insert the whole enzymatic pathway? or sell "glowing plant fuel"? How do you get around the fact that this uses quite a bit of the plant's energy? Are you planning on attaching promoters that hack into the plant's day/night cycle? or seasonal changes?

For what it's worth, it's trivially easy to grow your own bioluminescent fungi:

http://en.wikipedia.org/wiki/List_of_bioluminescent_fungi

They don't last as long as the plants, but if you get the right ones then you can eat them once you're done.

Hey - if you are in the Bay Area, I'd like to see them in person. simone brunozzi at gggmmail.

This is pretty cool tech! I kicked in quite a while ago, and am still waiting for my seeds. While I grumbled a bit, I was given a chance to vote for spending more time for higher quality--which I did. I am looking forward to when I have more time to work on things like this.

One of the things YC has encouraged us to do is ship early, so we are now going to ship twice to our Kickstarter backers - you'll get an early prototype and then version 2.0 as well. Looking forward to sharing our progress with you!

Sweet! I just bought 5 acres in South Florida where I plan to have a combo art center/research playhouse with greenhouses and different solar systems. The glow in the dark plants will be a nice accent.

What role did the DIYbio community play in fertilizing your business? What role do you see that community playing in future 'non-trad' bio-tech start-ups?

The founders met at Biocurious, a DIY Bio space in Sunnyvale, and many of our early backers came from that community. It's also a key influence behind us offering the DNA open source and behind offering the maker kit so that others can do their own plant engineering. DIY Bio means many things to many people, eg education, community, fun, but I personally see it as the foundation for entrepreneurs launching such businesses.

I'm interested to see the reactions this will garner from the anti-GMO crowd.

The usual suspects are of course against this, but one of the reasons we started the company was to change public perception around GMOs. The technology has enormous potential to help the world, but it's being held back by public opinion. What's frustrating is that this opinion is basically not based on science or evidence, but is mobilized because people are against things like biotech patents and large companies dominating the food chain. The glowing plant changes that discussion and we hope that people develop a more nuanced and balanced perspective once they have a GMO in their own home.

I can definitely see the potential for greatness that GMOs have. I'm really glad that you're trying to change the misguided public perception around them.

Plus it is harder to tell your kids they can't have glow in the dark plants than GMO food.

Just bought a seed packet. Very cool!

Will the plants produce seeds? Reproduce? Will the glow genes reappears in successive generations?

Thank you! The seeds are fertile and will retain the glow for a few generations at least (eventually evolution will turn it off due to the metabolic cost on the plant).

Unless you selectively cultivate them for luminescence, right?

I funded this project on kickstart. Can't wait to get the seeds.

I fear I know the answer to this already but just on the off chance -- is there any possibility these might be sold in Europe in the foreseeable future?

Foreseeable future yes, near future no I'm afraid. EU regulations require monitoring of every one of our plants which works for GMO's in fields (but is costly!) but is untenable for a consumer product. I'm optimistic the rules will change one day!

Do mean that EU regulations require you to monitor the plants once you sell them? Possibly to prevent an enterprising young man from planting 1,000s of these in the woods behind his house in order to have the coolest rave ever?

That's exactly what they require unfortunately. You have to know where they are planted and monitor the other plants in the area around them, obviously impossible for a consumer product!

Kickstarter backer here. Have you considered teaming up with another KS project (Click and Grow)? They've got a neat "self-watering" plant setup:

http://www.clickandgrow.com/smartpot

This seems a bit gimmicky, how does Glowing Plant thnk it's business will fare compared to Glofish?

http://www.glofish.com/

We are in touch with the Glofish founders, their business is doing well and they are in thousands of retails stores. The biggest difference though is that we have a very strong technical team who can continue to develop and improve our products, Glofish licenses from an academic research group. Like a software business this is a rapidly evolving technology so its important to have technical co-founders - YC obviously understands this!

I'd like to see this in a plant like a climbing rose. It would be very neat to cover a trellis, gazebo, or something similar!

I think people are more sensitive to having GM animals, which is kind of irrational when compared to the hideous monsters selective breeding has generated in goldfish.

I'd never have a glofish but I'd buy a plant in a heartbeat. I'd guess you'd sell a lot of these through retail on impulse, just like carnivorous plants - kids get dragged to the DIY superstore and walk away with a pacifying plant.

This is really cool.

I can't wait to see what the next several decades bring in this space.

It's all fun and games until a pet decides to eat it and gets sick or worse.

Reply on news.ycombinator.com