The Climate Biotech Podcast

What if Therapeutic-Grade Biotech Was Used for Environmental Remediation with Pranam Chatterjee

• Homeworld Collective

Share episode

Imagine proteins engineered to seek out and bind toxic heavy metals, cleaning up contaminated sites and potentially treating metal poisoning in humans. 

In this episode, Duke University professor and entrepreneur Pranam Chatterjee shares how his has developed two impressive AI tools transforming this field: MetaLATTE, which predicts whether proteins will bind specific metals, and the upcoming MetaLORIAN, which generates custom peptides designed to target particular metals like cadmium, lead, or copper. These technologies represent a significant advancement over traditional remediation approaches, potentially offering more precise, selective methods for environmental cleanup.

What makes this work particularly exciting is its dual potential—the same protein engineering techniques could address environmental pollution while simultaneously developing therapies for human metal poisoning. From brownfield remediation to industrial metal recycling and medical applications, these programmable proteins could offer unprecedented flexibility in how we tackle toxic metal contamination.

Visit chatterjeelab.com or huggingface.com/chatterjeelab to explore these tools yourself and see firsthand how AI-driven protein engineering is revolutionizing environmental remediation.

Send us a text

[00:00:00] Pranam Chatterjee: we think our technology can be applied to many different environmental questions and, through the students

we were able to crowdsource with my lab ways to really have an impact and it became very clear that. An important problem was the sequestration or removal of heavy metals. Heavy metals is more than just an environmental problem. It's also a health problem too. So really this problem became one of our most impactful questions in our lab. 

[00:00:33] Daniel Goodwin: Welcome to the climate biotech podcast, where we explore the most important problems at the intersection of climate and biology, and most importantly, how we can solve them. I'm Dan Goodwin, a technologist who spent years transitioning from software and neuroscience to a career in climate biotechnology.

As your host, I will interview our sector's most creative voices from scientists and entrepreneurs to policymakers and investors. 

[00:00:59] Dan Goodwin: I am thrilled to have this conversation with Pernam Chatterjee Pernam and I actually did our PhDs next to each other at MIT, where he was a young rising star and surprised absolutely nobody that he went off to go be very quickly a professor at Duke University.

He's a young dynamic, not only professor and academic, but also founder of two startups. Or maybe three, I think we'll learn about that here. Pernam was part of the genetic engineering revolution that swept through biotech as I saw it in about the 2015 to 2022 era, and did some really amazing work on the protein engineering side of Cas nine.

So one of the things we're really excited to talk about is, what we learned there, what kind of problems? We're surfaced at that point and how we can think about the field wide productivity we saw in pro engineering and genetic engineering in that era and what that might mean in the future.

Also, PNA is just up to very cool things in his lab, and so I hope not only talk about some of the cutting edge of biotech, but also talking about some of the cutting edge of climate and see specifically where those two interact we're very proud here at Home World to support Pernam as part of the Garden Grants program, and I'm really excited for everyone to hear about the work that he was doing with protein engineering and metal binding.

But before we go too far into the geeky weeds, we gotta be rooted in who we are as humans. Pernam, who are you? Where did you grow up?

[00:02:16] Pranam Chatterjee: Yeah. Hi Dan. Thanks for having me. I'm really excited to be here. The journey here has been pretty interesting and taken a lot of different detours, but I. Am originally from Georgia where I was born and raised and, got really excited about a lot of things from not just science, but I was also interested in what people believe and how they acted.

After doing school in Georgia, I went to Dartmouth College and New Hampshire where I was. Studying religion for a few years prior to transferring to MIT where I met you for both, all my undergrad masters and my PhD. So it was a nice little detour, but now I'm really excited about all the cool things that you talked about and excited to talk about it some more.

[00:02:58] Dan Goodwin: Wait, I know you as a computer science and bio wizard. I didn't know you also had a background in religion.

[00:03:05] Pranam Chatterjee: Correct.

[00:03:06] Dan Goodwin: Could we hover about what? 

[00:03:07] Pranam Chatterjee: Yeah, of course. So when I was younger, I lived in the, on the Alabama, Georgia border. So people had very different maybe viewpoints than me about the world, and all I wanted to do was understand it. So the way I did that was to study and read different religious texts about other people's belief systems.

I even spent a summer in a Buddhist monastery in Atlanta. Just to learn more about different religions, read a lot of books, and then I felt that the natural next step was to study academic religion. And at Dartmouth it was a great program to do but, after studying a lot of different faiths, which was really exciting to me, I felt that my passion was in understanding the world through the lens of ob like scientific reasoning.

And the transfer to MIT was in that realm.

[00:03:51] Dan Goodwin: That's amazing. I mean, there's a parallel here with Francis Collins who put his faith very forward as he did cutting edge bio research. And I loved his point that they're not at odds,

[00:04:01] Pranam Chatterjee: No they're not. And personally Dan, I'm not a religious person myself, but I really respect people who have different faith systems. I just really like to study how people think and why they believe what they do and, learn more about that. It's a great way to create connection community as well, which I appreciate.

[00:04:18] Dan Goodwin: I just can't help but riff on this a little bit. 'cause when I did my PhD in neuroscience right next to you. I would always think that you pick your scale on the brain, right? Are you like, are you a whole brain person? Are you a circuits person? And for me, I just went drilling all the way down.

I thought I was gonna be a whole brain person. I ended up as a subcellular person. And I feel like the religious angle is the same way. It's kinda like it's a meta brainin angle and then you drilled all the way down.

[00:04:41] Pranam Chatterjee: Yeah. And actually it's funny because I've always wanted to understand things at a very large level, and then I realized, to understand that I wanted to understand it at the molecular level of biology, but you needed a good tool to be able to really tinker there. When I was at MIT and I started there, I thought that first, interdisciplinary area to tinker with biology was material science, but actually spent a few semesters in the material science department. But then I realized, hey, there's a good chance that computer science and AI may have a big impact on biology in the future. Let me use that as my tool to discover and engineer new things in biology.

[00:05:15] Dan Goodwin: So now I'm trying to imagine first or second year PhD student PNO just had been studying religion. Did you know that you were always gonna be a biotech professor working at this intersection of human environmental health? 

[00:05:28] Pranam Chatterjee: I think so actually when I was at. Dartmouth, I really did believe in science and engineering as the way to solve a lot of great world problems. And so I knew actually I wanted to be a professor of bioengineering. I was very confident that was what I was going to end up doing.

And when I moved to MIT, I finished undergrad there. I had, started studying computer science and bioengineering and biology classes that it all came together. And so I've really been on this mission saying, Hey, I wanna be able to study what I wanna study likely now in bioengineering.

But this is the, this is the path that took me there. And by the time I was a second year PhD student, and you met me, Dan, I think I was already pretty focused on making this happen.

[00:06:07] Dan Goodwin: Yeah, and I think it'd be really interesting for people to know about your PhD work because I can just say from the outside looking in is that this was, say the 20 15, 20 20 era. It was a very exciting. exciting time in biotech because CRISPR was out now. It was the ification race, right?

Make CRISPR more accessible, make it more usable and that took us into the also cutting edge world of protein engineering. And that's where I think your PhD was, and you did awesome work. So I think it'd be a good foundation for this conversation to talk through it.

[00:06:37] Pranam Chatterjee: I can be very brief, but I really think you described it well. It was an exciting time for us and CRISPR we all know, is a very powerful tool that allows us to edit the genomes of, plants, of animals, of humans. When I came in, it was 20 15, 16, right when I started my PhD and I was really, I wanted to see, hey.

What area of CRISPR needed this introduction of computer science and protein engineering mixed together? And one of the problems that we saw was this inability to actually access a lot of the genome, right? We were limited to about 10%. Even though CRISPR was amazing, we could only edit about 10%.

And that was because of the way the CRISPR enzyme worked. And so I was like we gotta either engineer, discover, play around with these enzymes to be able to expand that reach. And that's what we did, right? Combine computer science and algorithms. You remember Noah was a fellow PhD student, Noah Giacomo.

He and I were like let's, we are computer scientists. Let's apply algorithms to discover and engineer new enzymes that could reach more DNA. And over the years we just. Discovered and engineered new and new enzymes. And now, in the first year of my lab at Duke, we actually have an enzyme that is pretty far reaching, I would say close to a hundred percent in terms of the DNA.

We can access a fully, what we call Pam Lis Cas nine. And that's really exciting because now Dan, that allows us to do a lot of things. Edit the mutations that cause genetic disease go into animals and or plants and edit genes that can control biofuel or bioethanol production and other applications.

So while I don't think I do too much CRISPR anymore, because I think we did our little bit, but we created tools I think that will have a lot of impact personally.

[00:08:22] Dan Goodwin: What we do a lot at Home World is we think about field building, and I think about, your work is very emblematic of a person doing great work in a well-built field, right? Like the, I don't know how you chose that problem of optimizing proteins for crispr, but it was a very, there's a very big gestalt at that moment, which is that, hey, yeah, only the PAMs that we can do are only 10% of the genome.

So it felt like there was this race. Then there was a lot of money and a lot of support and then a lot that brought in a lot of good people. I'm curious like kinda the experience of the choice of the problem and being in a field that was moving that fast. 

[00:08:55] Pranam Chatterjee: I think the good thing is that for me, I always wanna position myself where the work that I do could, people would use. As engineers, we want people to like, I. Take our stuff and make great things with it. And when you look at fields that are moving fast and moving and are actually pushing in a very impactful direction, CRISPR had a very clear direction that it was following.

You realize that this is a good way for your technology to have an impact. And so when Joe Jacobson, my PhD advisor asked me, pronou, what do you wanna do? You're a computer science undergrad. You had this weird background in religion what do you wanna do? You're at the media lab, you can do anything.

And I'm like CRISPR seems the right place to be simply because it, if you solve problems with genome editing, you can then do a lot of things with that, right? It wasn't just solving a prominent has a dead end solving gene editing. You can now edit genomes of all these organisms. You can be in climate tech, you can be in medi medicine and biotech.

You have all this impact. And so it was a very clear application of what I was trying to do, and I think that's why choosing these fields matters. However, Dan, I do think in the future we have to, we kinda just jump in and join a field. I think it's important to think about what new fields need to be produced for us to solve.

The problems that you guys are solving at the home world collective.

[00:10:16] Dan Goodwin: And I can't wait to get to your work on heavy metals. But we also get to step through a little bit on your journey, because you didn't go straight to becoming a professor, but basically straight. You did a short postdoc started a company. I'm sure people would be very interested.

[00:10:31] Pranam Chatterjee: Yeah, for sure. I think there was a little bit of a detour, so we did crispr, which is awesome. What the CRISPR story taught us was that good algorithms can help us discover proteins that have impactful outcomes, right? CRISPR Cas nine enzymes were one of those. Now. I said, let me take that same mentality and apply it to something different.

In my postdoc. I went to George Church's lab at Harvard, the question I wanted to solve was, what is another really exciting problem that could benefit from this paradigm? And one of the things we wanted to do was regenerative medicine and in fact, actually creating new cell types from stem cells.

I started a company amido, with my co-founder raised, $3 million, put that money into George's lab so that I could do my postdoc on developing. Cyte like cells from stem cells, these could be used for infertility and IVF and we did the work with great collaborators.

Put those patents back in Gamino and Gino's just had a first baby born a few months ago. we are actually the first US company to achieve phase three with the stem cell product. So all that, because I had this really little silly idea that let me develop an algorithm to discover transcription factor proteins that could make O cytes from stem cells.

So yeah, another example of that paradigm, leading to pretty impactful outcomes.

[00:11:50] Dan Goodwin: I'd say so. Gosh. I heard it for her first. 

Thank you for sharing. 

[00:11:54] Pranam Chatterjee: Yeah, the baby was Born and her name's Mia. She's in Peru, 

[00:11:57] Dan Goodwin: I'm really glad that you mentioned the funding story there because a big thing. The drives. Why we do our work at Home World Collective is to teach tools of empowerment to bio technologists, and especially now with a changing environment.

It's so important to know these recipes or know how to fundraise on your own, tell a story that's, backable and what you did for the cto. You said you stepped over it very quickly, but that was a very elegant, clever idea, which is you went out and you raised your own money to be a postdoc.

[00:12:29] Pranam Chatterjee: Yes,

[00:12:29] Dan Goodwin: To then do work that ended up becoming a startup. I mean, it sounds so obvious when you hear it, but first that requires a pi like George to say, this is awesome. Let's go. I can swing for it. But also there's a lot of very subtle sophistication on your end to say, this is the way I would do a postdoc.

[00:12:44] Pranam Chatterjee: Absolutely. I think I'm a person who wants to solve good problems and interesting problems that may not be. Classically fundable at the beginning of the story, right? Otherwise, I would have to operate in a regime where I am already an expert in something I wanna do and I need to build that expertise.

And, I think places like venture or investors like, angel investors can look at a person like me and say, pna can prob maybe solve this. Weird problem that he wants to solve. Let's give him a chance. And so I think, again, a lot of this was thank to my co-founders too at gto, who again believed in my science and allowed us to build this.

But that mechanism, I think is a really unique way of getting money for an idea that may need a little more added, really what you guys did right with Garden grants, I think is a very. I think a little cleaner maybe way than the way I did it because there, there is, in the end of the day I own the company that is funding my postdoc in a weird way.

So there was always some hoops to step through, but we did it and now we're better for it I think.

[00:13:46] Dan Goodwin: I think it's very clever. I really enjoyed watching this, and so congratulations, first of all. That's definitely something worth celebrating. And so you had a successful start of a company postdoc, and then you could have gone anywhere and you chose Duke. 

[00:13:58] Pranam Chatterjee: and that was a really exciting opportunity for me, right? Duke is an amazing place that actually supports interdisciplinary research very well. Two things you need to think about when you want to start. Obviously we had the opportunity to go into gato and.

Help run it or the next company. But I think what made the option of going to academia very attractive was that academia for me, serves as this incubator of really exciting ideas and it's structured. You have access to great students. If you're at Duke or MIT, or George at Harvard, you have access to students who wanna build things, but they want some structure in that process.

So for me, choosing a place like Duke was. Looking at how great the students were. Are people supportive? Are they collaborative? Can we do big projects like I did with George? What I did with Joe at MIT and the answer so far has been yes, because not only have we done the metal binders, which I think is one of the best and most exciting projects from our lab, but I do think that it's allowed us to build this new field, which I haven't even chatted about, which was this idea of.

Doing gene editing, but doing it for proteins, and other substrates like metals and other things. And so really excited and I'm glad you gave me that opportunity.

[00:15:10] Dan Goodwin: We're gonna get to that. I think it is just worth spending one more little back and forth on the lab itself. You've had the background in genetic engineering, and now like whole cell engineering, I would say maybe with the trend . And now you're doing prot editing 

so tell us the PNO lab has been around for a year,

[00:15:28] Pranam Chatterjee: Yeah, about two years. Now we're getting to that milestone, but really it's actually the integration of, remember that concept of algorithms to proteins that worked for both the gene editing, the cell engineering, and we were like what is the next way that we can apply? Especially the fact that these algorithms that we are developing are getting better and better.

We are developing new generative language models and really complex algorithms to design these molecules better. Can that be impactful? And I'll say this, Dan, half the lab are pure AI scientists. The other half are pure experimentalists and they learn to work together because they are very good at what they do.

So for us, in drug discovery, at least as the first. Instance, really it doesn't matter. Editing the DNA, it doesn't actually matter. Editing, RNA or what matters is what happens at the protein level. And so for us, we were like, wait, why are we editing DNA when I could just go to the protein and make the fix there?

And so we've built a system where. By using AI to design the things that bind to the protein that actually hit the protein, the bad protein, we're then able to not only bind, but get, remove it or bind and, make more of it or edit it in some way to make it healthy again. That is the idea inspiration was actually because of the algorithms, because of our knowledge of crispr, we were able to translate it to proteins and actually apply it to metals and other, and hopefully to like PFAS and other pollutant chemicals and elements that we think will benefit from this technology.

[00:17:01] Dan Goodwin: And this is the exact vision that took us into what became home world, right? How do you get the absolute cutting edge of biotech into these sort of deployments and you can totally sympathize with why things have been the way they have, is that it's so much easier to create a genetic engineering therapeutics company, right? If you can show the FDA that you can cure one person, they say, great, you can cure a million. And it's cost insensitive, right? And so for better or worse, that's the way bio has worked to date and there's been lots of money in it. And I get so excited when I see.

People like you that have shown the recipe of success in that to then take a step back and say, okay, cool. Can we repeat that recipe? But now in places where the market is a little less cut and dry like it is in the FDA, are you working on now with heavy metals?

[00:17:50] Pranam Chatterjee: That's a great point. And I wanna preface that saying that, I mean, we are almost as excited about the applications and climate and in the environment and others. It's, it is exactly that point. You have to demonstrate your technology in an arena where people already, very excited about the technology that's there, right?

Like the applications, the biology, curing disease. There's money there. People need new tools and we can deploy it. But that doesn't mean that our tools can't be applied to heavy metals or others. For us, this is such an exciting opportunity. I remember you and I had a great conversation saying, we think our technology can be applied to many different environmental questions and, through the students, right?

We were able to crowdsource with my lab like ways to really have an impact and it became very clear that. An important problem was the sequestration or removal of heavy metals. One thing I will say is that heavy metals is more than just an environmental problem. It's also a health problem too. Heavy metals are in our bodies so really this problem became one of our most impactful questions in our lab.

Can we design binders, peptides, which we are become experts at AI for peptide design. Can we design peptides that bind to heavy metals, either sequestered them in a field, right? Remove them from water and other contaminated environments, but also potentially could serve as a therapeutic as well to clear it in the body.

So it became like this one, like Swiss army knife type project that we became super excited about. And was led by an incredible graduate student who I had the fortune of bringing over from Singapore, who was doing a PhD at Duke and us in duke affiliated university in Singapore, brought her into our lab and she's really just become super obsessed with this problem.

She's a brilliant algorithm designer, and she's partnered with an undergrad experimentalist who's brilliant, who's been developing the. Experimental setup to not only design the binders to these, to cadmium, to lead to copper. You can imagine all these heavy metals that are pollutants, but also could be useful for recycling and apply them into an in vitro and potentially in the field environment to see if they actually have utility and purpose.

We've made so much progress just with the support of the garden grants, and we're excited to take that forward.

[00:20:16] Dan Goodwin: So everyone agrees that getting these medals out is. Good. And we can talk about humans and environment but why is it hard?

[00:20:24] Pranam Chatterjee: Yeah. It's, you know why it's hard from a scientific and engineering perspective is that,. It's not clear, you know, when we say DNA, right? Go back to crispr to target and bind DNA, you need to find the complimentary RNA sequence, right? That's your guide. It's, there's, the code is written in front of you for proteins.

It's not as easy, right? And that's what we do in our lab is saying, here's a cancer causing protein. It looks really wiggly. How do I bind to it? No one has figured it out, but there is data. There are other proteins that interact with these proteins. There is natural, like data from experimentalists who have collected all the protein or peptide protein interactions.

So at least we can learn from that for metals. The problem is not only is it not this classic like, oh, the complimentary sequence to the DNA. We don't have a lot of data for metal protein binding, and so now we're in this catch 22, right? We have this, basically this dilemma where we have none of neither of that, but heavy metals are, there's not many of them.

There's not like a billion proteins, right? Or a million proteins. There's a set number of them. We don't have much data for it, and so therefore we had to be really creative. In first understanding how proteins or peptides interact with metals and then using that information to design new ones. So that was the challenge, Dan, and I think, we've really bootstrapped a lot of different methods to figure out how to solve this problem.

[00:21:53] Dan Goodwin: So is it right to think of it as a A 3D electrostatics problem?

[00:21:58] Pranam Chatterjee: So you're talking to a person who doesn't think like that, right? I'm very much a high dimensional space thinker from a computational angle, meaning I think we can de deconvolute or um, simplify the problem into. Electrostatics or confirmations or wear positioning of the metal. But in reality, this is a very complex interaction that these proteins make and the idea of what we do, right?

We are a sequence modeling, language modeling. People think about chat, GBT, right? You trained on, I don't know, millions and millions or trillions of text examples. You didn't tell it that. Sentence was happy, and that sentence, would tells me about this. You just gave it. And so for the metal problem we just said, let the model understand that interaction, right?

Give it a lot of examples of proteins and then give it a lot of examples of proteins. Buying the metals, as many as we have, which did not too many, but enough to train this model and let the model come up with the million, billion variables. That it learned. And maybe, I don't know if I can interpret that in terms of electrostatics or, and I don't really need to, but I know that the model has understood it because we've tested the model out in practice.

So that's how I think about it. And I know a lot of structural biologists and like classical chemists, biophysicists look at me and say, oh, he just abstracts everything. But in reality, I know that these interactions are complicated and it would be. Unfair for me to try to pinpoint metal protein interaction on five, six properties.

In reality, the answer is probably a hundred million.

[00:23:36] Dan Goodwin: So the takeaway here is, your endpoint is not a pie mall thing with a little atom floating around. Your endpoint is an assay and you're either binding or not binding.

[00:23:45] Pranam Chatterjee: Correct. And I think there's a lot I. Involved in that and the way to learn it is to let the data, do the talking and intrinsically we call the self supervised learning, learn about the interaction itself. Until we have the information to do the designing.

[00:24:03] Dan Goodwin: This work in your lab turned into a paper called meta Latte. 

Am I right? 

[00:24:08] Pranam Chatterjee: yeah, it's I don't even know what latte stands for. Metal binding prediction with latent embeddings. Latte. And it's come even further. So Metal Latte was our first model that was great to just. Tell you, Hey, here's a protein, which heavy metal or any heavy metal does this protein bind to?

It's actually available online. You can type in any sequence of any protein and it'll tell you, oh, it binds to copper, or it binds to zinc, or it doesn't bind to any metal, and it'll just give you a check mark based beside the heavy metal that you want.

[00:24:38] Dan Goodwin: Do you know why that's really awesome is that we just went to the. Moss landing site with a battery fire where it was nickel, cadmium batteries.

So we'll get the sequencing. 

[00:24:46] Pranam Chatterjee: Right, 

And you may wanna know if there are, organisms either that have proteins that do that, maybe that's gonna be interesting to see if they evolved it. Another thing is, you may need proteins that do bind to zinc, cadmium, other heavy metals that were contaminated.

But here's the thing, Dan. Just predicting whether a protein binds to a metal is not enough. What we wanted to do was actually generate new binders to different heavy metals like vanadium or bled or mercury. That would be important. Or even, copper or cadmium. And so to do that, we said, okay, we have metal latte, right?

That predicts now we have another model that can generate new. Peptides and proteins I've gotta make, put these two together. I've gotta let the generator, the thing that makes new molecules, new peptides, decide what to make based on what metal latte says, right? If you make a peptide and it does not bind to your heavy metal, you should not make that peptide.

And so now we are able to guide the generator. Via metal latte to create a peptide that binds to cadmium, to copper, to cobalt. And the funny thing, we call this algorithm delorian. I know it's kind of a silly 

name, but, 

yeah, it's super nerdy name. But but basically met Allan's goal is to create, you give it a medal.

I need to bind to it will use metal latte to then generate that binder.

[00:26:13] Dan Goodwin: Cool. We're gonna talk about this next year, and you're gonna have a humanoid robot then taking Meor and doing it. So you've got such a great track record of deploying things. I'm gonna ask you the hard question which is let's say you get, you create the perfect metal binder. Who cares?

[00:26:26] Pranam Chatterjee: No, of course. I think we have a lot of stakeholders in this problem. Of course, we've been, did our due diligence, reached out to industry and there are companies who are really excited and who's tasked with. Sequestering heavy metals in polluted environments.

There are green tech companies that are working on this problem, so definitely I know that there's interest, but more importantly, there are a, incredibly large number of contaminated sites around the world. Both in developed and in developing nations where people who live in those sites are continuously being exposed to these heavy metals, and there are health consequences, there's consequences for the water, for the ecosystem, for the environment, and so governments.

NGOs and also companies in those areas that are interested in dredging out, right? These heavy metals. This is your ticket, right? No need to have these large mechanical devices that are, dredging. And my dad used to work in mechanical dredging of wastewater. Now if you have a biologic that is scalable, producible and deployable, you are going to be able to do this much better.

And the thing is, Dan. This is not just cleaning things up. There is also this idea of recycling metals for environmental and industrial applications like people doing silicon chips, not saying silicon, but I'm talking about the other metals involved in the process of creating like Nvidia chips or other chips.

This is a way to recycle those heavy metals, right? Get more and so there are stakeholders in various different arenas that can benefit from this. Our goal is to make the. Idea reality and then go out and say, Hey, I got you a very selective binder for copper. I have this very selective binder for cadmium.

Like what is your purpose for it? And we'll work with them to do that.

[00:28:12] Dan Goodwin: In your shoes, I would see two different use cases or ways to think about the service that proteins can do. One would be. The sorbent. So you have a really good binder that can programmable bind and unbind which is very similar to how we think about, say, carbon capture. The other way is chelation.

So you make the, you can toggle through, knocking off an electron here or there. It's either soluble or it's insoluble. Do you have a thesis on one or the other right now?

[00:28:38] Pranam Chatterjee: Yeah, I think we're very focused on chelation, especially divalent metal ions that are predominantly in these contaminated environments that, it's really that, that is causing this like, charge state, right? That is causing these metals to acquire their, harmful properties and may, toxic properties.

And so I think that's the focus first. Just being able to pull away metal ions and sequester them where they can then be chelated or then be, sequestered or removed or destroyed in some way. That I think would be a step that we could take and help people with.

[00:29:10] Dan Goodwin: It's a hard question but what is the relationship between binding and chelation?

[00:29:15] Pranam Chatterjee: Yeah it's something that we haven't figured out. This is I like to make the analogy back to what we do on the protein side, right? We don't just make binders. We actually attach the binders to an enzyme domain that actually removes. Ubiquitin or add phosphorylate, a phosphate group, right?

Not just binding, but editing. And if you think of metals, it's the same idea. What do you edit on a metal? You remove electrons, right? Like you're basically altering the properties of that. So my thought is this, Stan, I think that binding may not be just the only thing you need to do, but binding is the hardest thing to do, right?

Once you bind, now you have, you are there. You're at the battle, right? You've gotten to there. What will it take for us to actually improve the chelation? That I think is something we can think about. Whether it requires another enzyme, chelator, another protein, we can then play around with that. But you don't have any chance of even studying that problem or like solving that problem without the binder that hits the heavy metal first.

So I think we may think about it as a two part problem. Who knows? Maybe by interacting with it, we alter its chart state and we can chelate the metal ion better that way.

[00:30:23] Dan Goodwin: I mean, if you could have a binder that programmable unbinds, when the charge state is changed by the chelation, you have the magnet and it's like the catch and release.

[00:30:32] Pranam Chatterjee: Patch and release. Yeah. And that I think would be something we are trying to engineer into. And binders are, they don't attach forever. They have a KD and they, we are going to have to be able to play around with that. And that's why we're doing the experiments right now to really see what is metal Lian creating for cobalt for copper, and then play around with those properties later.

[00:30:51] Dan Goodwin: It's really exciting to me because it's a, first of all, I mean, pollution is a big personal area of mine. I think if we can reframe pollution as a pharma problem or a programmable problem, we get a lot of biotech talent, exactly like the PNO lab, right into this problem. And so I think you're really leading from the front.

The interesting thing is that the deployment is two totally different like environments, right? There's the mammalian system, humans. And then there's the really gross brownfields problems and I think both are extremely interesting and impactful in both ways. I'm just kinda curious where you at, do you have, is there a front runner for you or how do you think about, a protein that could go into two totally different systems?

[00:31:34] Pranam Chatterjee: I think it's something that is very difficult to model. And to understand. Because the thing is we, most of the time we're modeling interactions, protein metal. Or others , in a vacuum, right? This is the protein, this is the metal. That's it. These are the only things that are interacting.

 Basically, you've mentioned two different environmental settings that we would have to take into account. It's hard, and I think our goal now is to incorporate that information in our designs. This is gonna go into x. Environment, right? Can we say, now let's create a binder that's very optimized from a stability, from a interaction capability for that, and then a new binder for this.

And I think that. This is where, I'm a very sequence modeling guy, right? I love sequences because I think they carry information. I think this is where structure, like people who do like structure based modeling can actually falter here because when you model structure, you're really only looking at what's happening here, right?

Locally, right there. The cool thing about sequence is that information you can feed in a lot of functional information in, and that could be. Exactly the situation. You're suggesting more functional, more environmental questions and information that could be fed into the design.

[00:32:46] Dan Goodwin: I mean, the way you just set that up, it makes me think that you're thinking more like whole cell transcriptomics or like whole cell engineering rather than 

single small peptide binder.

[00:32:54] Pranam Chatterjee: And not even just wholesale. You can make the thing bigger, right? Like wholesale is great, but think about like whole environment, right? Aware. Metal binding design, right? Environmental way. I think that is very possible, and it's just how you feed in that information. We actually had a really great paper in January just for therapeutics, but the idea was not when you design a therapeutic peptide to a protein, not a metal, but a protein.

I need to incorporate. Solubility. I need to incorporate toxicity. I need to ex half life. These are bigger than just that interaction, and I think this is the same concept. And so this multi objective optimization problem is something we're going to have to incorporate in the metal binding question.

[00:33:36] Dan Goodwin: And people who are listening to this won't see the air quotes that you just put up,

[00:33:40] Pranam Chatterjee: Yeah. It's 

definitely like a multi objective optimization has been almost impossible in AI to this point. Remember Chachi, BT is zero objective. You just fill in the text. Dali image generation is one objective, right? Text guided image. You type in text, you get an image or video.

So the idea of multi objective, just from an AI perspective is tough, but for a problem like metal binding or other environmental questions. We're going to have to get good at that if AI is gonna have an impact.

[00:34:08] Dan Goodwin: All right. I'm gonna ask just a few more questions and we'll. What we can wrap up gracefully, because I mean, I love this conversation. I could talk about it for a long time. The the challenge I think with thinking about the human health and the environmental is that sadly it's gonna be very cost constrained, right?

So if you are making a chelation that the, therapy one, there's already something in the market. You need to show that you're better, made cheaper, et cetera. The other thing is that there's also a kinda to play with your hyperdimensional side of things. There's not really a good model, I would say, for toxicity of heavy metals.

And so you, we can go to the environmental side, but if first it's worth just hovering on that, which is that you've got two fantastic startups under your belt. And not that everything has to be a startup, but if there was to be something here, have you thought about what needs to be true or what these 

things might look like?

[00:34:55] Pranam Chatterjee: Of course. I think one thing to even to start translating, you really do have to, I know it's, you said either the health problem or the environmental problem. I think you have to identify that first. If the challenge is beating out another potential therapy to the clinic, that's a very different challenge than saying.

Let's figure out a way to get people and investors and venture capitalists excited about the climate, biotech problem. There are two unique things, and so I would approach the design and the deployment pitch from different angles for those two things. The problem is that we know we're in a very difficult time from a exploratory funding perspective, right?

Federal funding gives us that opportunity, so we need to start being creative. For me, it's what we talked about previously, Dan, which was. Who are the stakeholders, right? And who there are companies who wanna recycle heavy metal so they can use it in their industrial processes. I can talk to them, I can be like, I got you.

Here are your cadmium binders, right? You need this to be able to efficiently mass produce XX. That may not be the end goal for us, but you gotta take it in that step. And this is the same in biotech. I may wanna solve, I don't know, a rare pediatric disease that is really important to solve, but there's five kids.

But to do that, first I need to go and show the technology in, I don't know breast cancer or liver cancer. And once I get there, which is very important to solve, I have that opportunity to solve more niche problems. And so maybe that's how we have to do it. And formulate the problem that way, and I think it's gonna be a challenge, but we can do it.

And Home World gave us that opportunity to say, let's even start thinking about this problem.

[00:36:28] Dan Goodwin: About heavy metals and yeah. And it's one of those things where regulation has both a good and a bad connotation, right? And the FDA people argue that it's so expensive to create something new that things don't happen. But then once the FDA anoints it. It can be potentially deployed more broadly.

And I think with the environmental side, I see the same thing where if you can, what I've learned from, unofficial conversations with the EPA is that if you can show them that yours is a better filter, you can become the standard and then all future, cadium producing effluence now need to go through you.

[00:36:59] Pranam Chatterjee: That's a tough point. And I think that's, those are goals that people like me can set right. If I can become standard state of the art right in cadmium binding, that gives me an opening to have a disruption or to potentially get into the field. But you have to hold yourself to that standard.

We cannot just be like, I made a binder amazing. I need to be very good. And it needs to be adoptable by stakeholders and by people who would actually put them into practice.

[00:37:23] Dan Goodwin: Yeah. And it's what we see in the mining industry too is that there's this point where there's all these mines that have been frozen because they say if you take one atom off this site, you have now assumed the whole economic liability of this environmental disaster.

[00:37:39] Pranam Chatterjee: It's a risk. I mean, this is the point of. Company building or venture building or whatever, or even in our case, just project building, is that balancing that risk? I sound so like businesslike in this, but balancing that risk with the actual reward here is I think, I've gotten to learn this problem.

Many thanks to just going through the Garden Grants program. Learn the problem and realize that the risk is worth taking. It's a very important problem, but understanding that these are complications that could arise makes you design better one, but also makes you test and prototype better before you end up deploying it in the field.

[00:38:16] Dan Goodwin: I think it's so important. And for people to hear your recipes of success here, because sometimes in scientists we kind of like, ah, I'm gonna sound like a business guy for a second. 

But it's actually very powerful to at least be able to have that side.

[00:38:28] Pranam Chatterjee: And we're confident. And I think that's the thing is that I don't wanna tell people, oh, we can't do this, but maybe I can just, let me try. I I actually really think that it's good to project some confidence in the hard problems because otherwise people really do shy away from this problem.

I'm like, this is possible. We just, we have to work hard. It's tough, but we, if we throw, and that I think is. I don't see enough with my fellow scientists because they're just too afraid they're gonna be held to it. I'm like, let me be held to it. let my lab be the lab that is tasked with generating the next generation of heavy metal binders and sequesters and solutions.

And I think that's fine. I think we have a great shot at making it happen.

[00:39:05] Dan Goodwin: This has been so much fun. I'm gonna wrap this up with our four rapid fire questions at the

end, and then

Let people know where to listen to. More or to learn more about your lab. First question what is a single book paper art piece or I or idea that changed your mind and shaped how you develop as a scientist?

[00:39:23] Pranam Chatterjee: I think this is a very great question. It's actually a documentary I watched this documentary, I remember it about. A case of schizophrenia it sounds very medical, but in reality it just said, this is a problem that is both not understood and needs a solution.

So I was like, wow, being introduced to that type of problem was really inspiring to me and saying, maybe not a bad problem, but I wanna be that person who says poorly understood, but needs a solution. I. Let's create. So let's create tools to do that.

And that's why I take these big risks.

[00:39:54] Dan Goodwin: And that documentary is called A Case of Schizophrenia.

[00:39:57] Pranam Chatterjee: I forgot the name I think it was called Janie's story. It was about a little girl who had pediatric schizophrenia or juvenile schizophrenia, and it just it was, I don't know, it just really inspired me, just from a scientific perspective. I was very interested in studying belief systems and religion, which I think is still cool, but that was like wait.

 This is like. A problem someone needs to solve and there's others like this that I need to solve. Let me go and do that.

[00:40:22] Dan Goodwin: What is the best advice line that a mentor gave you?

[00:40:25] Pranam Chatterjee: Yeah. So George and Joe told me this. They said never very specific, never license your own technologies to someone else. Always make it happen yourself. And this is why I started two companies, right? Because I could have licensed the technology to a pharma or some other biotech.

But it was that because you can make something cool, you can license it, you get whatever resources from that, but they may just table it. And if you care about your solution, you need to. Show to the world that this is something that can work but by licensing and you let someone else do that.

And so for me, believing and championing your own technologies, your own ideas, was something that was taught by Joe and George. And they both told me that independently. And I really took that to heart and really worked on that myself.

 

[00:41:19] Dan Goodwin: I'm trying to stop myself from going on a riff about technology licensing offices. I'm 

[00:41:23] Pranam Chatterjee: I know. I know, I know, dad. I know you know it better than anyone.

No, your experience is amazing and all right, third out of four, if you had a magic wand to get more attention or resources into one part of biology, what would it be? 

I think it has to be as, I think climate has an audience. I really think it has to be the rare, like the n equals to rare disease. I know I've been a climate tech, but I really think that those people they'll ask you, they like, can you work on my. Can you work on my disease?

[00:41:54] Pranam Chatterjee: I went to a rare disease conference and they're like, can you work on my disease? I, there's two kids that have it. I'm like, I'd love to. I really wish someone would let me do that. But both in industry and in research, it is very difficult to get resources there. And I know we have, there are a lot of other problems to solve, but I really hope that, there will be more resources shifted in that direction.

And I hope that these problems can get some more recognition.

[00:42:20] Dan Goodwin: Right on. What's a skill that you think bio scientists need to invest more time into developing into themselves?

[00:42:27] Pranam Chatterjee: Yeah, I, it's easy for me to say ai, but it's not, I think a lot of my experimentalists are really good at at I. Not understanding ai, but they can interface with it. So I think the ability to quickly interface with new technology, right? Whether it is ai, whether it is quantum computing, whether it is high throughput screens, like the quicker you pick that up.

The further a biologist, a bioscientist will be able to move their research, always looking for the new things. It doesn't have to even be related to your field, but you're like I go over and I'm like, oh, look at how great. I don't know. Deep seek is working. I don't know, some AI tool. Let me immediately integrate that into my research.

Those people, those scientists, my students, hopefully they can do that. That's gonna propel their work so far because it introduces interdisciplinarity, but also gives them much more capability. 

[00:43:19] Dan Goodwin: This has been so much fun and I love those answers. As we wrap how can people find you and what would you want to draw their attention to? 

[00:43:26] Pranam Chatterjee: Absolutely first come to our website. There's two places I wanna go to our website, so chatterjee lab.com. And that, you can find my name and that'll be great. You can see the cool stuff my students do. I really wanna. Show you that this is all the students' work. I'm just their spokesperson.

If anything. Two go to our website. Go to hugging face.com/chatterjee lab. You're going to see that's our website where all these algorithms I've talked to you about. They're there. You can type in a protein and get heavy metal you can generate a new, like peptide, whatever you want.

And so I really encourage you guys to go there, but if you just wanna reach out to me, the best way is LinkedIn I think. So find me on LinkedIn, send me a message and I'm really happy to respond and converse about this work.

[00:44:10] Dan Goodwin: Chatterjee. It's been an absolute pleasure, my friend. Thank you,

[00:44:12] Pranam Chatterjee: Thanks Dan. Appreciate it.

[00:44:14] Daniel Goodwin: Thank you so much for tuning into this episode of the climate biotech podcast. We hope this has been educational, inspirational, and fun for you as you navigate your own journey and bring the best of biotech into planetary scale solutions, we'll be back with another one soon.

And in the meantime, stay in touch with homeworld on LinkedIn, Twitter, or blue sky. Links are all in the show notes. Huge thanks to our producer, Dave Clark, and operations lead Paul Himmelstein for making these episodes happen.

Catch you on the next one.