AI, Cancer Vaccines, and the Future of Personalized Medicine | Sid Sijbrandij on GitLab + Biotech

Sid Sijbrandij (0:00)

The trials are not meant to cure you. The trials are meant to get maximum data. Maybe sometimes as a side effect, they cure someone, but they're not designed for that. These companies, they pay a lot of money, and they need to get the maximum amount of data per patient. So that means only having them on one drug at a time, because otherwise how you're going to attribute the effect. That makes sense for these companies, but it doesn't make a ton of sense for the patient. And I was not able to participate in any trials because I wanted to do multiple things, and I didn't want to take a three-month pause and then only do one drug, because it was very likely to end up taking a drug that doesn't work in a trial. And even worse, you might even be in the placebo part.

Xinghui Yin (0:37)

This week, Mike and Devon sit down with Sid Sijbrandij, co-founder of GitLab and a cancer patient who has gone deep into the science of his own treatment. Today we discuss mRNA cancer vaccines, how they work and what it takes to produce them for individual patients. Sid talks about his experience going founder mode on his own cancer, from experimental treatments to open sourcing his medical data for researchers, and how he's working to create a world where personalized cancer treatments can become more effective and more accessible for everyone.

Please welcome Sid Sijbrandij.

Misha Shalaginov (1:10)

So Sid, just walk us through first principles, like soup to nuts. How do mRNA vaccines work and like why do they work? And what goes wrong in producing them? Like why don't we have an mRNA vaccine yet for like every disease and so on?

Sid Sijbrandij (1:22)

Yeah, it's a magical technology. In mRNA vaccines, you get to encode about 1,500, 2,000 base pairs. There's some starting and ending sequence, but in between, you're able to encode about 15 to 20 antigens, things that your immune system from then on will be on the lookout for.

You want to encode for things that are unique to your cancer. You don't want to attack healthy tissue. We also like stuff that's expressed in the cancer, that's easy to recognize, that's stable. But there's lots of factors to it. It's not a binary thing. This is good, this is not good. It is incredibly hard to make them. First of all, mRNA, it quickly breaks down. So you have to enclose them in liquid nanoparticles. Even that is quite hard. But then to design it, the more effort you put into the design and the more expertise, probably the better the end result is gonna be. And this is where like AI is gonna be incredibly helpful. But right now, we're missing the training data. We're just getting started. No one's published a lot of training data yet. So there's gonna be this quick ramp up. And for my second vaccine, I just said, you know what? We're just gonna talk to all the experts. So we talked to, I think, seven to 10 different companies. And Echo helped us organize kind of a seminar series to get all the input of all the experts and make a really good vaccine. What's also incredible is that it might not be that we just do one vaccine and then you're done.

It might actually be beneficial for patients to take another vaccine. Because the safety profile of the mRNA vaccines is incredibly good. So you might wanna kind of, for at least some time being, not for the rest of your life, but do a couple and make sure that the immune system is on top of things. At least that's the strategy I'm doing for myself right now.

And it's incredibly fun to kind of design it together and have all these different considerations like how, what is the data from the biopsy? Is this really something? Is this gonna have adverse events? What's the stability? Is this gonna be the winner drowning out the rest? Because there's a little bit of that. Like you stimulate the immune system in 15 ways, but only like six things really hit, in my case, which is typical. So the other ones are losers. What are we gonna do with the losers? Maybe create a vaccine with just the losers. Or maybe also say, hey, these winners, the immune system picked that up. We want to do that again because it was successful.

So, or maybe new things with new insights. So it's, nobody really knows yet.

But because of the safety profile, you can allow yourself to be on the leading edge of science. And it looks like for melanoma, they're incredibly effective. And now we'll see which other indications and what might be the cocktail we have to combine with the mRNA vaccines to make them work in tumors that are harder to turn hot.