BillionToOne Is Solving One of Biotech’s Hardest Problems

By Y Combinator

Categories: VC, Startup, Design

Summary

BillionToOne solved the 'needle in a haystack' problem of detecting one mutated base pair among 3 billion by adding synthetic DNA markers before amplification, then using machine learning to subtract noise—turning a biology problem into math. Two PhD students built a $4B company processing 600K+ tests annually in just 2 years.

Key Takeaways

1. Convert hard biology problems into mathematical ones: BillionToOne added known synthetic DNA molecules before amplification to quantify distortion, then used ML to subtract noise from sequencing data. This enables detection of signals 1 in a billion.
2. Start with first principles on the hardest constraint: The founders asked 'if we reduce noise, what conditions become detectable?' and realized cystic fibrosis and sickle cell (world's most common genetic disorders) became viable from maternal blood alone.
3. Pre-market validation creates expansion room: Their prenatal test achieved ~20% market share and 600K+ annual tests, giving them credibility and revenue to tackle the harder moonshot: early cancer detection before stage one.
4. Radical simplification beats incremental improvement: Before BillionToOne, genetic screening required invasive amniocentesis on high-risk pregnancies only. Non-invasive blood-based testing from maternal samples was considered impossible—reframing the problem unlocked adoption.
5. Speed through focused scope: Went from PhD idea to commercial test in 2 years by narrowing focus (maternal blood DNA first), not trying to solve cancer detection immediately. This allowed rapid iteration and early market entry.

Topics

• Noise Reduction in Signal Detection
• Synthetic Controls for Error Correction
• Prenatal Genetic Testing
• Minimal Residual Disease (MRD) Detection
• Biotech Product-Market Fit

Transcript Excerpt

One in 11 babies born in America this year will be screened by a genetic test that didn't exist a decade ago. Can you articulate like the needle in the haystack problem that you have to solve? There are 3 billion base pairs in the human genome. In a lot of the human diseases that we are detecting from mom's blood, cickle cell disease, cystic fibrosis, etc. It's usually only one base pair that's different. So you're looking for one base pair that's different out of billions. And that's where the ...