Ag Tech Talk Podcast: Lavie Bio Uses AI Engine to Save Time and Money Bringing Products to Market (Part 1)

Ag Tech Talk Podcast

Advertisement

Bringing new products to market is a costly and often quite lengthy process. Anything that can shorten that process saves time and money. Lavie Bio used MicroBoost AI to develop two new biological fungicides it plans to bring to market over the next couple of years. We talked with Russ Putland, Executive Vice President Commercial and General Manager USA for Lavie Bio, and Nir Arbel, Chief Product Officer at Evogene, Lavie Bio’s parent company, to learn how those solutions came into being and how the company plans to ensure regulatory approval and encourage end-user adoption.

Ag Tech Talk Podcast Ag Tech Talk Podcast

Podcast Transcript:

AgriBusiness Global: Let’s do a little background on the company. What is it you’re offering, what are you’re doing, and what separates you from other players in the market?

Top Articles
ADAMA Reports Fourth Quarter and Full Year 2023 Results

Russ Putland: Lavie Bio was born out of out of Evogene, which is our parent company who provides us. You know the technology and the AI, big data and that foundation.

We strive to improve food quality, sustainability, and agricultural productivity. And we do that through the microbiome, right through microbes. That’s the whole idea. And doing that with microbes, we have to leverage that technology to be able to do that. So that’s kind of the high-end view of what we try to do.

And really it’s about replacing synthetic products or enhancing food production to be healthier and sustainable. And to do those two things you’ve got to have productive agriculture, and you’ve got to have increased value.

You know it’s nice to have an altruistic view of the world where we’re going, you know. Make it a better environmental place, and we’re going save our soil. But in order to do that, if you don’t have financial sustainability underneath it all, it fails.

The products got to stand, and produce more, or produce better, or produce attributes. There’re a whole bunch of different ways you can do it, but at the end of the day, it has to be financially sustainable.

So, that’s what we strive to do is put those tools in our producers hands so that they can grow either more bushels, or more targeted bushels, or healthier bushels of grain or fruit, or vegetables depending on what market we’re in. And then, in turn, have the supply chain recognize that value all the way up to the consumer, where the consumers willing to pay for healthier, better, or increased value in the food.

ABG: Okay, regarding the microbiome. We’ve been hearing so much about that recently. I know we’ve written stories about it. Where are we in the process of understanding what’s there – all the different pieces parts that make up that microbiome?

RP: Yeah, it’s. It’s very different than what you know if you compare it to traditional synthetic chemistry and agriculture, or even human health. It’s very, very different in that. There’re billions of microbes in the world and in the synthetic world, we’ve run out of groups and active ingredients, and we haven’t really seen a bunch of innovation or new since the ’80s, to be honest.

And now, with being able to tap into these billions of microbes, it becomes more of a challenge to figure out which ones are beneficial.

They’re there, and we have these wonderful microbes that do great things, but we need to be able to see them clearly, so that we can actually make them living factories in a plant and use them for positive use in a natural way.

Nir Abel: If I may, chime in here. I think also the field of microbiome and microbes and understanding it – it’s quite nascent. It’s had a huge leap in the last 10 to 15 years.

But in terms of life science product development, that’s pretty much born yesterday. So, it’s a very exciting field. You see a lot of investors interested in the space. You see a lot of companies going into the space. But I think that the real magic of this field is yet to come, because there is so much more to learn, so much more to understand, and we are really just scratching the surface right now.

ABG: Maybe this is a good time, then, to talk about the solutions the company is offering. You mentioned MicroBoost AI, the tech engine. How does that play into how the microbiome in the products you’re trying to come up with?

NA: MicroBoost AI started out becoming an engine several years ago, when Evogene decided to use or leverage it’s a know-how in technology, in the genomic space. So the basis of micro boost. AI is the ability to understand genomics, but not as previously, originally, mostly at plants. Now, looking at the microbes and understanding the genomics of the microbes very well. The challenge, when you want to look at the microbial communities, is that there is so much information that we need to delve through to really understand what it is you’re looking at, what it is you’re trying to find, and it’s truly a needle in a haystack at any given time. So, what we are trying to do, coming up with our solutions, is to deconvolute those very challenging questions.

So, the basis of MicroBoost AI was to collect a very significant amount of genomes of microbial genomes and break those down into genes and those genes to in turn can understand what their functionalities are, and to group them together. So, we have, if not the biggest database, one of the biggest databases in the world for microbial genes and their functions, which is cataloged. So, we are able to cluster genes together towards understanding what their functionalities are, and that can be leveraged by Lavie, obviously, because when they’re they are looking to develop products and discover new microbes. They can utilize these catalogs to try and understand what sort of functionality they’re looking for. Where does this functionality exist and what bacteria are in their refrigerators or are in the public domain that they can reach out towards – they can find these certain functionalities. So that would be the biggest advantage of MicroBoost AI.

On top of that, what we can do by understanding the genomics is to run very sophisticated comparisons between different samples or different microbes, and what that allows the scientists to do is to compare and discover new functionalities and the genes related to those functionalities, meaning that if you know, and maybe this is a good example. So, Lavie Bio was looking for a long time at what genes would be related to shelf life, which is a non-obvious question. Right? So, bacteria evolution didn’t guide them towards shelf life, obviously, but they are at the end of the day, they do have some sort of shelf life as a product. So, what MicroBoost AI was able to do, for Lavie was to once looking at bacteria that had a good shelf-life, and those that hadn’t, was to highlight which genes are most likely to provide their microbes or their product with a good shelf life, and that has huge impact when you’re looking to develop the product.

RP: So, you think about the why; why does it matter? Well, it’s a matter of you need the microbes to stay alive while you’re using them, right, and to establish, and to create that living factory within a plant.

Also, from a commercial standpoint, it’s pretty nice to have it where you can keep the product for a year and still use it right from inventory management. There’s a commercial aspect to that, as well as not just the viability.

ABG: There are a lot of factors to consider when you’re developing a product. The development of traditional synthetic products can take a dozen years and hundreds of millions of dollars. And that’s the products that actually make it. And that’s doesn’t count all the things that are tried, get a year into it and figure out aren’t working. Does this system save you time? Does it save you money? There are billions of microbes out there. Trying to figure out how they all work together. The permutations are probably practically endless.

RP: I’ll maybe just walk through a high level of our process, and then let Nir speak to the technology underneath it. I think it would serve as well. So, it’s not different than synthetic chemistry, and you start with a product requirement. We have to know what we’re trying to build, why, we’re trying to build it.

We’ve got to think about what it’s going to cost to build us, and you know what the cogs are going to be to build it, and whether it is usable in today’s environment and regulations. So, all that piece gets what is the need and what does that product look like?

And the closer you get that at the start the better you’re going to be through the whole process.

Then it goes into Nir’s world where you know they’ve mapped these genes into functions, and you know that’s all being done and cataloged, right. And you know we’ve had a catalog that was called Taxon that we purchased. We’ve got some, you know, public domain microbes that we know a lot about, and then we’ve added to that.

Israel’s a great place to source microbes, because you have all different climates, right? So, you’ve got a whole bunch of diversity right in the country. So, it’s very fortunate for us. And as you map those you start to pick out the single microbes or the consortium of microbes that fits that product requirement. And then MicroBoost takes over. You start to pull from there and pull combinations. And I I’ve heard our science leaders say that you know there’s combinations come out as suggestions that this will work that traditional science just wouldn’t put together. It’s just not a reality.

NA: Russ spoke exactly to the point and this is something that’s crucial for people to understand that. The computational is not there to give out the final product. So, no one is expecting that we put information into the computer and some sort of AI would give out a single microbe that is, you know, the perfect solution. The whole point of what we’re trying to de-risk the entire process by identifying a fewer and more achievable amount of microbes that you can test, and that would give you the insight that you know it actually works and the prediction was good enough.

So, the way that efficiency and cost saving works is that once as Russ mentioned, you start out with the product requirements. You try to understand what it is you need. You can go to the catalogs and kind of pick out the genes that you need in advance. But usually what would happen is that there would still be some functionalities that you’re not absolutely sure what are the related genes that’s and then building out the correct experiments that would come from either the public domain from publications, or would come out, you know, from the experimental aspect of the company, and feeding them correctly into the algorithms that can then flesh out the highest rewarding genes for that functionality.

And the point at the end would be, is, you know, connecting all those dots together. Because if you want this in a frequency aspect, it’s, quite easy to find right, because you would. You would easily find bacteria that is anti-fungal, or entire insects, and so on. But then that does not fulfill all the product requirements. You need cogs as well. You need regulatory aspects as well and combining all of those together in a quite efficient and rapid pace – that is what allows the scientists to come out with what they feel is the best potential product to test that out. And that’s with the simple solution of coming up with only one bacteria product.

Once you want to look for more than one bacteria to build a consortium. Then everything I said gets, you know, exponentially more complicated, and it’s a necessity to use computational tools here, because teaming bacteria is definitely not straightforward. So, figuring out how we would work together is a very challenging question. There are multiple solutions out there, but using computational tools is, is very, very effective towards that.

So, combining all these attributes together is key. But then really testing them out and making sure that it works is what really saves cost and time. And maybe as the last point that, like I mentioned previously, looking at shelf life is also key. And that’s definitely a time and cost saving, because the attrition rate for bacteria that have poor shelf life is around 30%. Shelf like is usually the last thing that you test because you go through the entire process and develop your product. And then the last thing you’ll do is grow it and then put it on the shelf and test it every few months to make sure that you know it’s sustainable and has a long shelf life. And if you figure out six months later then you have to go back and with a 30% attrition rate that means three or four out of every 10 products in development would fail at that point.

So, by being able to predict what are the genes that would contribute to shelf life, and scanning the bacteria from the get go, and making sure that they contain these genes, would really reduce the risk of failing at that time, and that would save a lot of time.

Hide picture