One person built this, on purpose.
PharmTwin is built and run by Andrew Albert — a solo founder with a genomics and entrepreneurship background. It exists because the gap between "pharmacogenomic science" and "what you can actually act on" was indefensible.
The problem this exists to fix
About 70% of US adults take prescription medication. 30–50% of them carry at least one pharmacogene variant that meaningfully alters how they metabolize or respond to common drugs — antidepressants, blood thinners, painkillers, statins, anesthetics. CPIC, the FDA, and DPWG have all published peer-reviewed dosing guidance for around 200 medications. The biology has been settled for years.
The interpretation has historically been gatekept by clinical pharmacogenomic labs (GeneSight, Genomind, OneOme) charging $300–600 per report and requiring a physician order. Meanwhile, ~40 million Americans already have raw DNA files sitting in their email from 23andMe or AncestryDNA. About 80% of pharmacogenomically actionable variants are already in those files. Those customers paid to generate the data. They just don't know there's a drug-response report inside it.
PharmTwin is the report. Same source data PharmCAT runs at clinical sites, same CPIC and DPWG rules, same level of scientific rigor — just delivered directly to the person whose genome it is, for a price that doesn't require an insurance fight.
What's intentionally narrow about this
PharmTwin does pharmacogenomics. It does not do cancer risk prediction. It does not do polygenic scores. It does not do ancestry, carrier status, or any of the other genomic interpretation categories. That's a deliberate choice — pharmacogenomics is the one category where the FDA has already endorsed the underlying science (their Table of Pharmacogenomic Associations is the federal acknowledgment), CPIC has done the clinical heavy lifting, and the regulatory surface is small if we frame outputs as educational and "share with your prescriber."
Narrow is on purpose. It lets one person build a real product at a real price for a real audience instead of building a worse version of 23andMe.
Why a single founder, why now
The whole pipeline — variant calling, guideline matching, plain-English narration, conversational Q&A, doctor-handoff PDF — can be built on top of open-source tools (PharmCAT from Stanford, CPIC's published tables) and a small Python + Flask stack. Compute per user is pennies. No wet lab, no inventory, no shipping, no clinical staff. The right business form for this problem is a small one.
We don't sell your data. We don't have an advertising business model. We don't have outside investors pushing us toward growth-at-all-costs. The product makes money if the report is useful enough that you tell a friend.
Get in touch
Questions, feedback, prescriber inquiries, partnership ideas: hello@pharmtwin.com.