Scientists Have 'Hacked Photosynthesis' in Search of More Productive Crops | Hacker News Reader

Scientists Have 'Hacked Photosynthesis' in Search of More Productive Crops

npr.org

128 pointsevo_982 comments

I wonder - once we start replacing more extensive segments that represent complex functions, not only does it become a less probable natural solution but to steal a term from ML - perhaps there is a danger of temporal "over-fitting"... The natural version maybe sub-optimal in the current climate, but that historical baggage could have evolved to that state because it allowed the continuation of that species through a more diverse range of climates, something that could be useful in the future.

There is the potential that forcing a "more efficient" solution for _now_, could be also creating a more fragile and endemic species? Perhaps not even for the distant future, if details in the original in-fact make it better able to cope with subtle climatic variations more efficiently.

Crop plants and livestock animals are already pretty much unable to survive in the wild, outside of farms. So the "overfitting" is as old as farming.

The dangers of monoculture as well.

True, but this is direct, cultivation is by proxy, and does not necessarily remove unexpressed genes.

That's a confusing thing to say... surely GMO doesn't necessarily remove unexpressed genes either?

To provide a counterargument, with GMO we additionally have a decent understanding of the underlying mechanics and which genes have changed. This puts us in a better position to correct things if they do turn out to have unintended consequences.

> That's a confusing thing to say... surely GMO doesn't necessarily remove unexpressed genes either?

You may be correct, this was speculation based on imagination on my part - I do not know if unexpressed genes were removed in this experiment, I only have a suspicion they may have been part of the "inefficient" original sequence that was replaced.

Livestock animals live in the wild pretty well. Google “feral hogs.”

Not quite true. Some livestock animals do ok in the wild: hogs and, in some warmer climates, cattle. There are feral goats in the Caribbean. But the climate has to be right. You don't see feral goats in Maine. Or even Ohio.

Domesticated chickens, OTOH, are toast anywhere there are predators. There are feral chickens in Hawaii. Nowhere else that I know of. There are also few feral sheep, as they need to be sheared.

Domesticated horses also tend to not do well. They've been selectively bred for traits that make them somewhat fragile without specific care.

As for crops: most domesticated food crops don't tend to become invasive. Every once in a while you'll see volunteer corn left over from last years' crop. and volunteer tomatoes are practically ubiquitous in backyard gardens. But more and more, the seeds from food crops are not fertile, they're hybrids. And even in backyard gardens of "heritage" crops, without human intervention the weeds almost always win out in the end. The only vegetable I've ever seen that persisted long after the gardener was gone was, oddly, rhubarb.

That's a far cry from OP's comment that livestock animals are "pretty much unable to survive in the wild."

I agree with you that crop plants are much more fragile and unable to compete without watering and fertilization. Tomatoes bred to make big 1 lb bags of water for their seeds will have a rough time compared to the tiny tomatoes made by the wild ancestor of tomatoes.

https://www.smithsonianmag.com/travel/why-wild-tiny-pimp-tom...

I'd argue that this is partly a philosophical argument: what do we mean when we say “the wild”? None of the feral animals listed, with the possible exception of feral pigs, could survive if humans hadn't killed off the major predator species. There are wild cattle where I grew up, cows that escaped one of the local farms. But there are no wolves or mountain lions, which in a true natural state would, I suspect, would take them down in short order.

So to be generous to the OP, he might have a point.

this is over... simplifying

That baggage may be more useful insofar as it allows species to adapt more quickly over generations to new climates. In contrast, the techniques we develop based on our intelligence can be put into effect much more quickly, effectively making us more agile in the face of change.

That said, monocultures increase our vulnerability to a class of threats even faster-evolving than our technology: diseases. As with all things, we should approach this with balance.

If you're concerned about this you should financially support some seed banks. They effectively keep backups.

Indeed, they are worthy cause. However I am not concerned, just interested. I've also visited kew and donated to their seed bank.

I think it's important to try to reason about the evolution of the things we attempt to modify but is not necessarily a cause for alarm so long as one modified species is not so aggressive as to risk completely supplanting other variations.

I think the right way to think about it is that modern enzymes are at least partly optimized to handle a wide range of conditions (more or less what you said), but also lock in a fair amount of legacy details that would be hard to evolve out due to the amount of dependencies on them (everything in biology is tightly related, even modular systems leak).

I think of it in an RL context- evolution has acted as a subtle force on the fitness, in which natural selection evaluates mutation- but the resulting actions and policies are extremely subtle and integrate a wide range of conditions, and weave a large number of players together. Then scientists come along an re-optimize for a single variable and claim they've "fixed a bug". No, it wasn't a bug, you just don't understand the full aspects of the biology.

(note: I work in ML and biology full time, but what I say above is effectively non-scientific, impossible to prove, and mainly for enjoyable thought-discussions).

> but also lock in a fair amount of legacy details that would be hard to evolve out due to the amount of dependencies on them (everything in biology is tightly related, even modular systems leak).

I remember visiting a clone of the Encephalartos_woodii (AKA "the loneliest plant in the world") which is thought to have evolved into a genetic dead end due to it's massive genome which is said to be 20 times the size of the human genome... I wondered if this was due to some kind of dependence, which sounds a lot like what you are saying?

it's not clear what the plants that have super-large genomes are doing. Some of them are perfectly capable and show little to no negative issues associated with their genomes. I think the case you're describing, there was just a large full-genome duplication that was tolerated (somehow the gene regulation still works and the appropriate amount of protein products are expressed, etc).

What I meant more is that most enzymes work in regulated webs and you can't easily tweak one part of the web without having knock-on effects on all the other genes. Nature seems to handle this via gene duplication- once a gene is duplicated, it seems like one can carry on the original function for its dependents, while another is freer to evolve new functions.

Sorry I don't have a better explanation for this- to really talk about this you kind of need to have a huge background and know a ton of biology and have read all the debates from all the players about fitness and evolution and even then, nobody really has good answers for this.

> Nature seems to handle this via gene duplication- once a gene is duplicated, it seems like one can carry on the original function for its dependents, while another is freer to evolve new functions.

I think this is a pretty decent explanation for the layman :) I might not appreciate all of the technical details but it's easy to imagine why this would work conceptually. There are plenty of computery analogies, like running a legacy program in production while developing an experimental one.

Maximally efficient is minimally robust.

there's a lot of baggage with photosynthesis because plants have been around for a long, long time.

in fact, they have been around and have existed in so many climates and microclimates that there are actually different types of photosynthesis: CAM, C2, C3, and C4. some came about because the gases in atmosphere changed over time, others because of heat and humidity.

if you have a half dozen or so houseplants, most likely you have plants that create food via 2 different types of photosynthesis.

for example cacti use CAM photosynthesis to prevent water loss during the day when it's really hot out. a number of succulents also use this and it's such an important part of survival in specific climates that it's actually an example of convergent evolution where something like the euphorbia plants in Africa developed while never sharing a close genetic ancestor with the cacti and succulents of North and Central America.

From a past life as a biologist looking at photosynthesis, IIRC, at least the photosynthetic systems of algae were optimized for (1) "dynamic range" of sunlight and (2) molecular stability.

(1) Intermittent flecks of light - not guaranteed continuous sunlight in the wild.

(2) every photon captured destabilizes the molecular assembly and some captured energy may be better lost than build up oxidative stress.

Of course, this is from a ten year old memory and may be completely wrong.

>"But it has what we like to call one fatal flaw," Cavanagh continues. Unfortunately, Rubisco isn't picky enough about what it grabs from the air. It also picks up oxygen. "When it does that, it makes a toxic compound, so the plant has to detoxify it."

Can someone explain this, the article doesn't go into much detail. What is this toxic compound, and why must the plant make it?

edit: Explained in detail in this article: https://arstechnica.com/science/2019/01/re-engineering-photo...

The semi-technical "Perspective" column

http://science.sciencemag.org/content/363/6422/32

in the same issue of Science is also excellent while still being more accessible than the main technical article.

http://science.sciencemag.org/content/363/6422/eaat9077

The article focuses on GMO food production which is great. But since this process is said to be more efficient than the natural one, could this be also used for consuming the excess carbon dioxide in our air?

I mean, in the sense that any plant that grows consumes carbon dioxide as long as we don't turn around and burn it, yes. Biohacking any particular species to express a particular trait isn't exactly solved yet, though.

I've long wondered if we could do to blue-green algae (Cyanobacteria) what we've done to Penicillium fungi, mutating like crazy, to optimize for CO2 absorbtion, then bury the mass.

> bury the mass

Or burn it - at least it would be a sustainable energy source.

Create a carbon cycle similar to the water cycle then?

Already exists. We've now identified many biogeochemical cycles.

An incomplete list:

https://en.wikipedia.org/wiki/Carbon_cycle

https://en.wikipedia.org/wiki/Oxygen_cycle

https://en.wikipedia.org/wiki/Nitrogen_cycle

https://en.wikipedia.org/wiki/Calcium_cycle

https://en.wikipedia.org/wiki/Silica_cycle

https://en.wikipedia.org/wiki/Phosphorus_cycle

https://en.wikipedia.org/wiki/Sulfur_cycle

https://en.wikipedia.org/wiki/Selenium_cycle

https://en.wikipedia.org/wiki/Mercury_cycle

https://en.wikipedia.org/wiki/Hydrogen_cycle

https://en.wikipedia.org/wiki/Rock_cycle

https://en.wikipedia.org/wiki/Carbonate-silicate_cycle

Easier to just burry it, treating it for use in combustion systems wouldn't be worth it given the energy density is low. Better off using Sabatier process with excess solar at that point.

You can also feed it to people and livestock. I've been growing Spirulina, a cyanobacteria, in aquarium tanks in my window for eating purposes. It's quite high in protein, antioxidants and beneficial oils.

When I get the time, I'll likely also try making some biodiesel and/or gasifying some of it to see how feasible it is as a fuel source.

Here's the question: how does it taste?

Fresh, it's pretty tasteless, and quite easy to incorporate raw into other foods such as a smoothy, soup, pesto, or a spread in sandwiches.

You should also look at Azolla. It grows crazy fast and is also high protein and fixes both carbon and nitrogen.

https://en.wikipedia.org/wiki/Azolla

Thanks, I'll see if I can find some.

I'm currently also experimenting with duckweed growing in my backyard, although I find that it doesn't play well with the other aquatic animals and plants, as it appears to smother them out.

If you have problems with duckweed smothering, azolla will probably be worse. Duckweed has to pull nitrogen from the water, azolla harvests it directly from the air. It makes a good feedstock and fertilizer though.

To make a differenece in CO2 consumtion, a large area has to be planted. On a global scale, think Sahara-area size. At such scales, the limiting factor would be the nutrient availibility in soil. So, such GMO might play a role; albeit not as the main factor.

To add: the change might play a role jn smaller emissions from normal agriculture.

This is correct, but it's also accurate to say that land-management will need to play a role in mitigating climate change (to the extent that's still possible): https://news.berkeley.edu/2018/08/29/improving-soil-quality-...

More efficient might mean less CO2 consumed to simply produce an identical plant. Would have to trial it. Assume this genetic manipulation might even stunt growth in some species. Again would need trials.

I cannot find the particular article I was looking for regarding withered leaves that plagued a long-term (10+ years) genetic manipulation project in sugar cane, but here is a list of problems I found: http://natureinstitute.org/nontarget/browse_titles.htm

From my observation of friends who were doing genetic manipulation of plants, it is not easy.

From the article:

> And they created super tobacco plants. "They grew faster, and they grew up to 40 percent bigger" than normal tobacco plants, Cavanagh says. These measurements were done both in greenhouses and open-air field plots.

The mass of the plant is proportional to the number of Carbons in it, and the Carbons come from CO2, so a 40% of mass increase means something like a 40% increase of CO2 sequestration.

[The Carbon content of the plant varies form plant to plant and in each part of the plant, and due to other reasons. The plants are made of Carbon, Hidrogen, Oxigen, and some minor but important atoms like Nitrogen, Phosphorus, ... The backbone of the molecules are made of Carbon, with some Oxygen and Hydrogen on the sides and some Nitrogen sparked here and there. So the proportion of Carbon is in a narrow range. The main exception is water, but the plant doesn't want to be too dry or too diluted internally, so also this is a narrow band.]

Yes, impressive results, but do we really want more tobacco plants? I think the appropriate test will be if the results are transferable to other species. And which ones would make the most sense? Algae, lettuce, spinach, weeds? Weeds make some sort of sense because they are already so resilient, often being the best-adapted plants in a particular environment. But do we want weeds that grow even larger and faster? Scientific research comes with a responsibility to try and anticipate and avoid any unwanted consequences.

For instance: Where are the insects? Scientists are trying to figure out what has caused the recent dramatic (80%) decline in insect numbers. One possible culprit may be a newer class of insecticide:

"Of particular concern are neonicotinoids, neurotoxins that were thought to affect only treated crops but turned out to accumulate in the landscape and to be consumed by all kinds of nontargeted bugs." https://www.nytimes.com/2018/11/27/magazine/insect-apocalyps...

Oops, the danger of unintended consequences.

The article mentions this. The technology was proven on tobacco, only because it is a "model organism" for plants. Researchers are attempting to replicate the results in cowpea, as it is a staple food crop in sub-saharan Africa.

This tech is explicitly being developed for increased crop yields. The results will be transferable to nearly all plant life, as the photosynthesis proteins are the most common proteins in the known universe.

We are refactoring code that is billions of years old, full of evolutionary cruft, to remove the technical debt.

> a 40% of mass increase means something like a 40% increase of CO2 sequestration

The carbon in annual plants is not considered "sequestered". Most of it will be released after harvest, either when the end product is consumed or the non-product organic waste degrades. Agricultural sequestration is usually talked about in terms of getting carbon into the soil and leaving it there. This tends to be done through promoting soil health, composting, selecting plants with deep root systems, perennials over annuals, etc.

https://en.wikipedia.org/wiki/Carbon_sequestration#Agricultu...

I agree, but I think that indirectly there is an increase. From the same article in Wikipedia:

> The decreasing of SOC content can be counteracted by increasing the carbon input, this can be done with several strategies, e.g. leave harvest residues on the field [...]

Yes, have a look at the work of Erb, Schwander, and others at the Max Planck institute on the CETCH cycle, for example.

http://science.sciencemag.org/content/354/6314/900.full

A field crop planted annually that grows faster or bigger isn't likely to make much of a difference to atmospheric CO2 as it will just release it again after harvest. Perennial plants (those that don't need to be planted each year) are better at sequestering CO2 in the soil, along with other benefits. The Land Institute does work on this, including developing (through traditional plant breeding) a perennial wheat variety which is being trialed [0].

There are other agricultural practices that promote CO2 sequestration in the soil. You might have a look at Project Drawdown, it ranks solutions to climate change based on impact and has a section on food [1] which includes agricultural solutions. For example, Silvopasture [1], ranked #9 overall amongst all solutions, involves integrating trees with livestock pasture. Another one is conservation agriculture (#16) which uses annual crops but promotes soil health in selection of species and method and timing of planting [3].

[0] https://landinstitute.org/our-work/perennial-crops/

[1] https://www.drawdown.org/solutions/food

[2] https://www.drawdown.org/solutions/food/silvopasture

[3] https://www.drawdown.org/solutions/food/conservation-agricul...

Why can’t you take all the leftover biomass after a harvest and sequester it? Then it wouldn’t “simply relaase it”. I’d image we only take a small part of the mass for food.

for which crop? most of the leftover biomass is used in typical crops.

For corn, the stalks are used as silage. For wheat, the stalks are used to make straw.

> For corn, the stalks are used as silage.

How common is this? In my area, the combines chew up the stalks and spit them out as chaff. I don't see many people collecting the chaff.

Not sure. The dairy farm I worked at used the corn stalks. They make a product called "silage" which can be used as winter feed for the cattle.

Curious. My neighbor was a dairy farmer. They did this too. No one else though.

It can depend on the price of fodder. If hay is cheap, there's not much incentive to spend time/fuel on baling stalks.

However, if they're not baled, they decompose in the field, which technically reduces fertilizer inputs.

I was under the impression that, in the case of corn at least, harvesting for silage or for grain was an either/or proposition. That is, if the crop is struggling and you don’t expect much yield as grain, then you might want to harvest it for silage to cut your losses. (I’m not a farmer, but I literally grew up in Podunk, Nebraska — a farm-adjacent village.)

That's what is known as 'feed corn' and is grown specifically as livestock feed. The specific strains of corn are optimized for producing maximum plant material, which is mostly stalk and leaves. A field of feed corn plants will usually be very tall, 8-15 feet. In contrast a field of sweet corn (grown for human consumption and possibly for processing into other foods), will be much shorter, since it's optimized to produce the most, largest and highest sugar content corn kernels, and any energy the plant spends to grow a tall stalk is wasted.

Feed corn is harvested with big combines that just cut down the whole plant at ground level and chop everything up. The results are piled up and left to ferment. Once fermented, you have silage.

I'd be surprised to learn anyone was harvesting sweet corn and saving the waste material to make silage. The equipment isn't designed to keep that stuff it just gets dumped back onto the field. And those strains of corn don't produce very much plant material since they're optimized for small plants and big cobs. Additionally you'd have to load all the waste material into trucks, which would significantly raise the cost of harvesting. And dumping that stuff back onto the field is a good thing, it helps keep the dirt down for the winter and decomposes into usable nutrients and fibrous material which helps reduce compacting, etc. It'd be expensive and labor intensive to try to capture the waste material from feed corn. It's easier and more economical to just plant feed corn or buy silage.

You're right about straw, though. The harvesters are specifically designed to leave the straw in row pile behind them. Then you run a baler over that and it leaves a row of straw bales in the field. Then you run a stacker over that (or a flatbed trailer and buck 'em by hand) and you have a haystack.

Pretty sure this is more or less what no-till farming does; hopefully someone more knowledgable can weigh in and confirm or correct me.

We haven't really figured out sequestration even from extremely CO2 rich sources like power plants. To do it w/ organic field waste you'd have to collect it all, send it somewhere and process it. In terms of sequestration bang for the buck (cost per ton) that sounds like it would make it low down on the list of priorities.

Actually we've know how to do carbon sequestration for plant matter for quite some time. The resulting biomass can be gasified and remaining carbon is turned into biochar which is inert in the soil for thousands of years. Not only is it a viable process, it's a fairly decent energy source which makes it cost productive.

This can be done using solar energy and a parabolic trough. Auger feeds ground biomass in one end, the other end gets biochar, water vapor and a range of fractions, from methane to tars. Methane and the tar can both be burned to power the auger and other processes, though the methane is easier to work with.

If you sequester biomass incorrectly, like burying it underground tightly with no oxygen, you could have further emissions due to microbes digesting plant matter into methane, which is a worse GHG than CO2

Replanting all the forests in the world that have been cut down will only replace all the firsts in the world. Eventually we run out of room for new forests.

The major battle in combating AGW is the quantity of carbon we have dug up from the ground and released into the environment.

I'd think bamboo production would be a good candidate for this

https://www.drawdown.org/solutions/land-use/bamboo

The individual plants are probably not enough. To really soak up some carbon at a noticeable level, you need to take a large amount of crop fields and have them go fallow, becoming forests.

This has been successfully demonstrated once before by G. Khan et al[0].

[0]https://www.theguardian.com/theguardian/2011/jan/26/genghis-...

I thought I had as well. But then you're talking about reengineering whole forests.

There’s a website for a project aiming to produce a modified version of rice using these improved photosynthesis pathways[1]. Some previous HN discussion on the subject can also be found[2]. Really interesting stuff!

[1] https://c4rice.com/

[2] https://news.ycombinator.com/item?id=8867365

I hate the writing in this. "This protein's super-important!" "Yay for Rubisco!"

Here https://arstechnica.com/science/2019/01/re-engineering-photo...

Thanks for sharing, this article is so much better than OP's. It answers a question I posted which the NPR article was very vague on, but Ars covers in detail.

the article is basically an edited transcript of the radio interview, where the quotes you mentioned feel much more natural.

While this is certainly interesting it doesn't have as wide an application as you might think. In tobacco the harvested product of value is in the leaves, bigger leaves means a bigger crop.

But take corn, larger leaves and a taller plant doesn't result in bigger ears of corn. A taller plant in general means that it is more likely to lodge or fall down in the fall before it's harvested. Some corn is harvested as silage meaning the entire plant is chopped and fed to cows. I've worked with farmers who experimented with varieties of corn that would be 3-4 feet taller than regular corn. The result was less feed value per acre than what they had been growing.

With soybeans, wheat or oats you can also likely get an increased amount of disease with more foliage. This development may see more application than just tobacco but it remains to be seen.

This is a modification which increases the efficiency with which the plants generate needed sugars which presumably are used everywhere in the plant, not just the leaves. Why would this only increase the size of leaves?

Indeed, if each photosynthesizing leaf becomes more efficient, wouldn't this decrease the number of leaves you need to support the non-leaf consumable part of the plant, allowing foliage to decrease?

In nature there is typically no free lunch. Optimizing this may lead to other issues such as lower drought tolerance of less natural pesticide resistance. Although as the author states, photosynthesis isn’t usually a rate limiting factor in nature so it could potentially be optimized.

It's neat to see this finally come to fruition (no pun intended). Someone has been anonymously posting about this research for the last few years on 4chan's science board.

"It's called Rubisco, which is short for an actual chemical name that's very long and hard to remember."

Anybody else wish he would've let us be the judge of that?

https://en.wikipedia.org/wiki/RuBisCO

Superweeds have been talked about before, in terms of weeds that are herbicide resistant. Kind of a yawn, really, unless you're a farmer.

But a weed that grows 40% faster than anything else out there? That's some scary stuff. (Palmer amaranth, for example)

Kudzu?

Anyway, my impression is that most plants bottleneck on water or nutrients before sunlight, so this probably won't make that big of an impact if it gets out in the wild, but it would be a problem on fertilized and irrigated fields.

Also for better solar batteries. Direct solar to dissociated hydrogen & oxygen. Might eventually be more efficient than solar generated electricity if you read some of the university press releases.

There isn’t enough information to assess why this is a good thing or if it will work.

What could possibly go wrong?!

A lot of things can go wrong for pretty much any significant technology.

Those technologies do require caution, and control, but I think this is still a good step forward, even if it ends up being unusable as is for whatever yet unforeseen reason.

Reminds me of a story from Tales from the Crypt, about a super-plant, an episode with Stephen King...

One thing that concerns me about hacking crops to have higher yields is that they may consume more soil nutrients, when we’re already in pretty serious danger of exhausting large patches of arable land. So then we’d need to use a lot more fertilizer, which is already an environmental disaster at its current scale with nitrogen/phosphate runoff.

There’s no such thing as a free lunch, unfortunately.

Reply on news.ycombinator.com