Australian ML engineer used ChatGPT, AlphaFold, and genetic algorithms to design a personalized mRNA vaccine for his dog's cancer

Fantastic. You're the man. We'll talk to you soon. Great to have a good one.

Take care.

Let me tell you about Reream. One live stream 30 plus destinations. If you want to multiream, go to reream.com. And I believe we have our next guest already in the reream reading room. Paul Cunningham is the dog healer.

Dog healer.

And now he's in the TV pin Ultra.

Paul, how are you doing?

How's it going, Jess? What's happening? It's going fantastic.

Great. I don't know if it's early or late for you, but thank you.

Or is it is it shockingly early?

The sun is just rising.

Okay. Just rising. Well, we appreciate you getting up early to come chat with us. Uh why why don't you uh take us through some of the some of your backstory, your history. I feel like you have a very interesting career that led up to this moment, and then we'll go into uh the actual uh story and process of of of what what went viral over the weekend.

Uh sure. Um, I've been doing machine learning since about 2009.

Mhm.

Um, went fulltime about 2015.

Mhm.

Uh, I ran the Sydney machine learning meetup group here for sixish years.

Um, uh, I worked with, uh, Shalto and Tristan on a robotic arm project.

Oh, yeah.

And, um, yeah. Uh, right now consulting.

Oh, yeah.

The Sydney Mafia. I forgot. Yeah, that makes sense that you did it over there, not not not you you weren't in the States for that project. That's fun.

Yeah, correct. Yeah,

that's great.

Very cool.

Uh and so yeah, t take us through the story of uh I I I actually lost this in the story like when did you when did you find out that your dog was suffering from cancer? What was the initial process? Was what at what point did you leave the traditional veterinarian system?

Um sure. So what what actually happened the pre-tory was Rose uh had some uh like skin rashes appear on on his skin.

Mhm.

And I took to the vet uh and he misdiagnosed it for for three times for about 11 months. So over a period of 11 months I took it to the vet was misdiagnosed and on the third time it started bleeding. So I decided to have the the the tumors removed and that's when it came back as cancer

unfortunately.

Um tried really hard to have additional surgery just to remove as much cancer as possible

and um to like essentially try kill it at the stem and um because it had been misdiagnosed for so long one of the tumors that got so large that wrapped around her leg and we weren't there's just not enough skin to like close it.

Oh. So that's when I kind of realized we needed to do try different options.

Sure.

Um and and then tried put on chemotherapy. Um

but uh none of the traditional stuff was essentially like like stopping it. It was continuing to grow.

Okay. Okay. And then so so when do you actually first go to AI tooling? Do you start at a very high level just sort of asking about dog cancer broadly? like I at what level did you come into the conversation with AI just understanding the capabilities?

Um well I I knew about uh Alpha Fold from um

the Alph Go days. So it was the progress the the progression technology

and um I just decided to chat tot one day in November 2024

like come up with a come up with a plan on how we can like potentially make a drug to block this cancer. I I didn't really know anything about cancer at this stage. I was just um

Yeah.

going through the process of trying to figure it all out. Yeah.

Yeah. And so uh what what happens next? Who do you actually call to? Because at some point uh you know it is just text in a box in an app or a website. Uh at what point do you need to go back into the real world to uh advance the next step? I imagine that chatbt at one point tells you like, okay, well, we'll need the DNA sequence and we can't get that just from a text box. So, where do you go next?

Yes. So, yeah. So, correct. So, then the the first actual piece of um of data we needed was the DNA sequencing.

Yeah.

And um yeah, uh CHBT recommended to reach out to Professor Martin at you know, it provided three other people, but it was like it gave all the reason that this is the reason why you should reach out to Professor Martin. That's really cool.

And through a mutual friend here in Sydney, I was connected to to Professor Martin and he was very receptive to to just taking it on. Um, extremely receptive. Yeah.

And so at some point you walk me through, you know, for those who are familiar with 23 and me. It's a saliva swab. What's the actual process for getting a dog's DNA sequenced? And then what's the file type that comes back? Do you just get a text file? So this is considerably more advanced than 23 and me. It is like we have I I have Rose's entire genome on my on a on an external hard drive I bought.

Wow.

So uh um the process to submit the RNA uh DNA sequencing was quite um cumbersome. So it was filling out spreadsheets and stuff to submit.

Um but what came back two weeks later was 300 gigabytes of data.

Wow.

Yeah. And had to push through that. Yeah.

And so and so at at this point you're not just dragging that file into a consumer chat bot at this at this point. You're you're starting to build custom pipelines. Correct.

Yeah. Correct. So again uh I use I use chatbt I use Gemini and I use Grock.

Yeah.

Constantly switching between the two. Mhm.

And um yeah, built out the pipeline to essentially go through the steps of um uh computational pipeline to get to uh um the mutation to sort of see what's causing the the cancer, the the root cause.

Mhm. And did you actually use alphafold? Is there like an open- source package that you can download and and run?

Yeah. Yeah, we we use alphafold too.

Okay. Um, so from the literature and from also additional LLM sessions, I I found out that there's a gene called seekit

that is one of the primary drivers for Ros's cancer.

Mhm.

And um what we what we essentially did was take uh her healthy DNA. So we we sequenced her healthy DNA

and we sequenced her um cancer DNA.

Mhm.

Compared them side by side. I brought like a genetic diff between the two.

Yeah. and um and then focused in on like the secret gene, pulled that out, model modeled it in alpha fold.

Okay.

And um and uh uh I used two different techniques to essentially look for drugs to um uh to try to block the cancer. One was genetic algorithms. So I ran genetic algorithms and we actually came up with a unique chemical compound that could block it. Mhm.

Um but the reason I didn't pursue that is because uh I actually talked to a chemist about having it made. But the problem with that is you have to like go through the the steps of uh you know um first doing it in like uh in in a test tube then moving to mouse models and moving on

further. So that's too complicated.

Yeah.

And um yeah the the other technique was docking. We docked a whole bunch of these chemical compounds called lians to the alpha fold 3D protein 3D structure of seekit and um mutated secret and um uh essentially discovered a drug that was very very strong at blocking it. But unfortunately the drug was is owned by a major US international uh um company. I I reached out to them for for compassionate use and

sure

they politely declined which is fair enough.

Okay. Um but that was kind of the there there's a second part of a full we used late in the pipeline but um that is kind of the start of the journey and uh uh around this around this time was about uh June 2025 and I through went through all of that and um it really took the wind out of my sales because I like I tried everything. I tried like to like see if I could synthesize it. I tried to see if I could like get hold of a pre-existing chemical and um yeah, one day I was walking Rosie down the street and I realized maybe I'm actually close to making a vaccine myself and got back on chat GPT and like I typed away and it said, "Yeah, you know, you're halfway there. You've already done the DNA sequencing. These are the next steps you need to do."

That's amazing. So, so, so back to the lab. You did mention we at this point. So I imagine you've looped in friends, colleagues, like who is around you on this project at this point?

Um, uh, it's myself and, uh, I run a small AI consulting firm here in Sydney.

Yeah.

Yeah. So just like just I kind of worked on in part-time for about two hours every day.

Wow.

Yeah.

It's remarkable. Um, so so so back to the lab and they wind up finally making the drug. Um yeah, so that was a a process in of itself. I went through and did the design of the vaccine construct

um and pushed uh I literally emailed it over to the mRNA institute at the at UNSW.

Yeah.

It was like um half a page of text.

Mhm. and and um the the the major blocker was actually getting an ethics approval

because you can't just like go and make a a mRNA vaccine in your garage like they don't let you do that in Australia.

No. Um, so, uh, I'd been notified that I had to create an ethics approval. And, um, again, I spent, uh, I don't know, that was three months of my life creating that. And it got to a point where we were actually going to have to modify, um, the university's license with the government because the vaccine was going to be administered offsite.

So, um, the ethics approval would have only been approved in June uh, this year.

So, uh,

Oh, wow. uh through a connection in in in America in um Seattle, I was connected to Professor Mary Maya. I don't know how to pronounce his surname

and uh she she is like the preeminent canine cancer person on planet Earth. Uh she connected me to someone in in in um a professor in Queensland, which is a state that's about a thousand kilometers. I'm not sure what that is in miles. uh uh north of here. I was talking chatting to her and then was just saying like I'm having trouble with eth ethics approvals and she uh said, "Oh, I actually have a I have an ethics approval with the government for that specific type of novel imunotherapies and um you know, I'm happy to take you under my wing." I just played it completely cool. I was like, "Oh yeah, cool." Um but actually like

you're jumping up and down. That's amazing. inside my head.

Remarkable. Oh, that's so cool.

So, um

yeah.

Yeah. So, uh once we got the green light to do that, uh uh it all sort of like lined up in in uh in parallel. I I drove Rose up to Queensland. Um we did the induction phase of the vaccine.

Mhm.

And then I just sort of waited to see the results.

And so, and um yeah, cancer is like a long fight, but it seems like there's at least some really positive signals. something like a 50% decrease in the size of the tumors. Is that roughly correct? Like how are you measuring progress these days?

Okay. Um there's been a lot of talk about that. So obviously the the visual um um the the the best uh like uh trait is the the reduction in the cancer size.

Yeah. Um we also took uh blood work which is going to be published in a paper later this year.

Mhm.

Um and uh just continuing to visually monitor her her tumors essentially.

Yeah, that's great. So where do you think this goes next? Obviously there's a lot of attention. Uh some people are saying oh maybe you'll launch a startup around this concept or try and democratize biotech further. Uh, do you want to just continue the story in some way or is it back to work as usual?

I think like um the process itself was uh way too hard and I think there's room to make it much much much uh easier for for not just people like me but everyone.

Yeah.

Um Yeah. So I think there's definitely room. I like I I even know I could probably do the pipeline now in maybe four to six weeks.

Okay.

Um and and that's important because the faster you can do the pipeline uh cancer is constantly mutating. So if you can run the pipeline f if you can outrun the speed of the mutation you can like essentially clamp it down.

Yeah. So uh talk about the long pole in the tent. Uh it it I imagine that just sequencing DNA takes time. uh actually producing a chemical or or synthesizing the actual vaccine, producing the product takes time. Uh it sounded like ethics waiverss and the and the approvals also took time, but some of those can be shortened, some of those are going to be harder to shorten. Where's the biggest opportunity that you see to speed up that cycle time?

Uh the computational pipeline itself can be sped up. M

um the sequencing can be sped up. Sequencing is getting better and better every 6 months. It's like on a double exponential.

Whoa. Um

I had no idea.

Uh uh uh and then there's we could probably do the the the vaccine itself manufacturer foster.

Mhm.

Um

and then ethics as well. I I think there's definitely that's probably the biggest room for improvement right there to be honest. like yeah I think for sure

and I think that's going to be a story that we'll see in a bunch of other use cases and categories where like the technolog is advancing faster than society and our legal system can can even adapt. But what I what I just love about the story is as amazing as as it is the role that Chad GBT and these other LLMs played in this process, it really is a story of your just like insane determination and agency and high effort over such a long period of time to uh to save your dog. And uh it's just incredibly admirable. I love it. And uh I hope that uh many many more people hear about this story. I'm sure you've been uh I'm sure some like

documentary crews and things like that have have uh reached out, but it's really

really special.

Well, thank you so much for taking the time.

Great to meet you. Keep keep us posted. Send us your progress when you when you put out the paper later this year. Come back on and we're uh sending our prayers to to Rosie.

Yes.

And uh it's you can see her.

Oh, there we go. Amazing. little air horn air horn maroon.

Well, thank you so much.

Incredible stuff

to come chat with us.

Yeah, great to meet you, Paul.

Have a great rest of your day.

Cheers.

Yes.

Cheers, Tom. Byebye.

Goodbye. Let me tell you about Shopify.