Codility VS DecodeMyBio

DecodeMyBio's answer:

People who already hold a consumer DNA file — 23andMe, AncestryDNA, MyHeritage, FamilyTreeDNA or Living DNA — and want to use it for medication decisions. That includes anyone starting a new medication who wants to check genetic compatibility, people dealing with unexplained side effects or a drug that simply did nothing, and those who prefer data-driven prescribing over trial and error. Many of them arrive holding a 23andMe export they downloaded before the company's bankruptcy and want a real home for it. Findings are written to be brought to a prescriber, with the gene, variant and published guideline shown for each one.

DecodeMyBio's answer:

DecodeMyBio focuses specifically on how your genes affect your response to medications, not generic trait reports. It reads a raw DNA file you already own — 23andMe, AncestryDNA, MyHeritage, FamilyTreeDNA or Living DNA — and answers questions about the drugs you take, tied to CPIC, DPWG and FDA sources across the 128 drug-gene pairs CPIC grades as actionable at Level A and B. Two things set it apart. First, results are an interactive dashboard with an AI assistant that cites its sources, not a static PDF. Second, it is explicit about the limits of your data: a consumer genotyping array cannot fully call CYP2D6, for example, and DecodeMyBio says so rather than guessing past the file. Access is a one-time $59 purchase you keep for life; the only recurring option is Decode+ Monitoring, which re-checks your saved medications when CPIC guidance changes.

DecodeMyBio's answer:

Most DNA upload tools lead with ancestry or broad wellness traits. DecodeMyBio is built around prescribing decisions: it tells you how you are likely to respond to specific medications, backed by CPIC, DPWG and FDA guidance rather than general associations. It works from a file you already own, so there is no new test and no wait. It is also honest about what your file cannot support — array data cannot fully call CYP2D6, and the product marks that rather than filling the gap with a guess. Pricing is a one-time $59 for lifetime access instead of a mandatory subscription; monitoring for guideline changes is an optional add-on. Your data stays encrypted at rest and in transit, is never sold or shared with insurers or employers, and you can delete your genetic data at any time.

DecodeMyBio's answer:

DecodeMyBio started from a simple observation: millions of people have a raw DNA file sitting in a downloads folder doing nothing, and the most useful thing in that file has almost nothing to do with ancestry. It is pharmacogenomics — how your body actually processes the drugs you are prescribed. That evidence is public and clinical-grade (CPIC, DPWG, FDA), yet almost nobody outside a hospital ever gets it applied to their own genotype.

The 23andMe bankruptcy made it urgent. The company was sold to the nonprofit TTAM Research Institute and AncestryHealth was retired, which left a lot of people holding an export from a service that no longer stands behind it. The file still works. It just needed somewhere to go.

The product was rebuilt once, deliberately. The first version sold static PDF reports, and that turned out to be the wrong shape: a PDF is frozen the moment it is generated, while the guidance built on top of your genotype is not. CPIC revises its recommendations, and an answer that was current when you downloaded it can quietly go stale. So DecodeMyBio became an interactive platform on top of an open genomics engine — you buy the interpretation once and keep it for life, and monitoring for guideline changes became a separate, optional layer rather than something bolted onto a document.

The other thing that shaped it was deciding to be honest about the limits. Consumer arrays cannot call everything — a full CYP2D6 read is the standard example — and it would have been easy to paper over that. Saying "your file cannot answer this, and here is why" turned out to be the most trusted part of the product.

DecodeMyBio's answer:

A Python service layer behind a modern TypeScript web front end, on managed cloud infrastructure with a managed relational database.

The part that actually matters is the genomics layer. It is built on open bioinformatics tooling that parses the raw exports consumer testing services produce and calls pharmacogenomic phenotypes from them, on top of published clinical references — CPIC, DPWG and FDA guidance, with ClinVar and gnomAD for variant significance and frequency. Those references are ingested as versioned releases rather than queried live, which is what makes it possible to re-run an old result against new guidance and say precisely what changed.

The assistant is a large language model, constrained to the user's own decoded data and required to cite the gene, variant and source behind anything it says.

Payments, transactional email and authentication use established third-party providers rather than anything home-grown.

DecodeMyBio's answer:

DecodeMyBio is direct-to-consumer, so there is no roster of company customers to list here. Its users are individuals who already hold a consumer DNA export and want it applied to their own prescriptions — typically someone starting a new medication, someone who has had an unexplained side effect or a drug that simply did nothing, and people who downloaded a 23andMe export before the bankruptcy and want it to stay useful.

We also do not publish customer names, and that is deliberate rather than a gap. These are people's genetic and medication records; a pharmacogenomics product should not be turning its users into a logo wall or a testimonial list. Data is never sold, and never shared with insurers or employers.