Why we are
building this.
A founder's letter from Marc van der Chijs, Co-founder & Chief Strategy Officer.
I've been thinking about longevity since at least 2017. Back then I wrote about it as an investor thesis — the idea that extending human healthspan would become one of the defining technological challenges of our generation. It was somewhat abstract at the time.
Then I lost two friends to cancer. That changed a lot for me, because something abstract suddenly became personal very fast.
When someone close to you dies from cancer, at some point — if you're the kind of person who builds things — you stop asking "why did this happen" and start asking "what is actually being done" and "where are the gaps." I started reading papers about cancer immunotherapy, personalized medicine, and neoantigen vaccines. And I found a gap that shouldn't be there.
The gap in the market
The standard tool used in clinical neoantigen prediction is pVACtools. It's open source, well-maintained, and widely used. But it ranks candidates primarily by one thing: how strongly the peptide binds to the patient's immune presentation molecules. It ignores whether the gene is actually turned on in the patient's tumor. It ignores whether the mutation makes the peptide look sufficiently foreign to the immune system.
The TESLA consortium — 36 research groups with experimentally confirmed immunogenicity labels — identified five key parameters governing neoantigen immunogenicity. Multiple academic groups have demonstrated that machine learning models incorporating these features outperform binding-only ranking. Yet in clinical practice, most vaccine trials still use binding affinity as the primary signal. I saw an opportunity to build something better.
What we learned building Elyra
Together with Sean Clark — my business partner for almost a decade, co-founder of Nasdaq-listed Hut 8 and a TSX-listed ETF — we founded Helixion Therapeutics in Vancouver. Sean is a builder. He is not a career academic, but someone with deep technical ability who can take a complex problem and ship a working solution. We complement each other well.
The most important finding during development was that gene expression is the single strongest predictor of neoantigen immunogenicity — even more important than binding affinity. It accounts for 36.3% of the model's predictive power. This makes biological sense: if the gene isn't active in the tumor, nothing downstream matters. But pVACtools doesn't use expression for ranking. That's the gap we exploit.
We also learned, the hard way, that training data distribution matters more than volume. An early model trained on 47,000 peptides — including infectious disease data — performed worse than one trained on 18,000 cancer-specific samples. Sometimes less data is better, if it's the right data.
The vision
The full vision is an end-to-end service: a research partner or biotech sends us tumor sequencing data, and we return a validated, ranked neoantigen report with structural analysis and expression profiling. Neoantigen prediction as a managed service — the computational biology department that small vaccine companies don't have and can't afford to build.
If any of this resonates, get in touch.
"Your immune system is designed to find and kill cancer cells. The question is: which mutations do you pick? I realised this is a computational problem — and that the current tools are not solving it well."
"Gene expression is the single strongest predictor of neoantigen immunogenicity, even more important than binding affinity. But pVACtools doesn't use expression for ranking. That's the gap we exploit."
"AI has changed what a small team can build. The bottleneck now is knowing what to build — the domain insight, the judgment about which features matter. That's what Sean and I bring."
We are looking for computational biologists, immunologists, and bioinformatics engineers who want to work on a problem that matters.
See Open Roles