如何定序自己的DNA？他用超廉價成本在家完成

Bradley Woolf, with a bioinformatics background, used a $7,500 Oxford Nanopore MinION sequencer at home, spending two months setting up the entire workflow to sequence his own genome from oral sampling, extraction, loading, to variant calling, doing it end-to-end five times. He also noted in his article that this is still far from the $2.7 billion and 13 years spent on the Human Genome Project, and the first run's results are not yet diagnostic grade for medical decisions.

(Previous context: What is decentralized science DeSci? How it works, pros and cons of introducing DAOs, potential tracks sorted) (Background supplement: OpenAI launches life science model GPT-Rosalind: three core capabilities in genomics, protein engineering, and chemistry)

Table of Contents

Toggle

• Six governments spent 13 years
• Extremely complex process
• The genome itself is not "magic"
• DNA is a stable reference, RNA is the current you

Key Takeaways

• Bradley Woolf, with a bioinformatics background, used a $7,500 Oxford Nanopore MinION sequencer to sequence his own genome five times at home, doing everything from oral sampling to variant calling himself.
• Compare to the Human Genome Project: $2.7 billion, 13 years, a joint effort by six governments—now one person can replicate it by spending two months setting up the equipment.
• He emphasizes that the genome itself is not magic, just a "reference layer"; it only becomes meaningful when queried against databases like VEP, ClinVar, and gnomAD. The first results are not yet diagnostic grade and cannot be used to make CRISPR edits based on a report.

He scrapes the inside of his cheek with a swab against the mucosa for sixty seconds. Then he drops the swab into a centrifuge tube with cold phosphate buffer, puts it in a used centrifuge, and spins at 2000 g. He takes the supernatant, adds lysis buffer and proteinase K, lets it sit to break down cell structures. This sounds like a high school biology lab experiment, but the end result of this process is a complete human genome sequence. This result was not produced by any lab—it was published by Bradley Woolf himself on his blog, and he did it five times at home.

Six governments spent 13 years

In 2003, the Human Genome Project was officially completed. It was an international project jointly funded by the governments of six countries: the United States, the United Kingdom, Japan, France, Germany, and China. It started in 1990, took a full 13 years, cost about $2.7 billion, and involved national-level sequencing centers and thousands of researchers.

Bradley Woolf has no lab, no government funding. He has a palm-sized Oxford Nanopore MinION sequencer costing $7,500, a laptop running MinKNOW software, a GPU for processing sequencing data, a vortex mixer, a heating block, and a used centrifuge bought from eBay. It took him about two months to assemble and tune the system until it could consistently produce high-quality results.

What six governments did in 13 years, he says he did five times at home—a comparison that surprised many.

Extremely complex process

Sampling is just the start. After cell lysis, Bradley Woolf uses the NEB Monarch high-molecular-weight DNA extraction kit with magnetic beads to pull DNA out of the protein debris, then measures concentration with a Qubit fluorometer to confirm how much DNA is available for sequencing. The standard protocol requires 1,000 ng of DNA; in one practice run, he only got 13.9 ng but still pushed through.

Next is DNA end repair and adapter ligation, using ONT's SQK-LSK114 ligation sequencing kit with the NEBNext Companion Module. Before loading, he checks how many active pores remain in the flow cell: over 1,200 is excellent, 800–1,200 is usable, 500–800 is borderline, below 500 means it's not worth running. After confirming active pores, he adds the sample, starts MinKNOW, selects high-accuracy or super-high-accuracy mode, and the machine begins reading DNA sequences in real time.

Sequencing is not the end. Raw signals must go through base calling with Dorado to convert to strings of A, T, C, G, then alignment to the GRCh38 human reference genome using minimap2 to find differences from the standard, and finally variant calling with Clair3. The entire workflow, from sampling to analysis, he does alone from start to finish at home.

• Core machine: Oxford Nanopore MinION sequencer, $7,500
• Peripheral equipment: Laptop for MinKNOW, GPU for Dorado base calling, vortex mixer ($50), heating block ($250), used centrifuge ($400)
• Per-run reagents: NEB Monarch DNA extraction kit, NEBNext Companion Module, ONT SQK-LSK114 kit, Qubit fluorometer, AMPure XP beads

The genome itself is not "magic"

Bradley Woolf says, "The genome itself is not magic; it's a reference layer."

What the sequencer spits out is just a long static string of letters; it tells you nothing by itself. The meaningful action happens after sequencing—running this sequence against databases like Ensembl VEP, ClinVar, gnomAD, and PharmGKB to find out whether a variant is associated with disease risk or whether a certain gene affects your metabolism of specific drugs. He himself checked PharmGKB to see how his drug responses might differ from the average person.

But he draws a clear line. The data from the first MinION run is not yet diagnostic grade; variants with insufficient depth cannot be trusted. He writes more directly: this is definitely not the kind of thing where "you use CRISPR to edit yourself because some AI said so."

Understanding your own DNA and making decisions based on it are two completely different things.

If you had your own complete genome report in front of you, would you dare to adjust your medications based on its suggestions, or would you first think about how trustworthy that report really is?

DNA is a stable reference, RNA is the current you

Bradley Woolf's ambition doesn't stop at sequencing. He says DNA is a stable reference, while RNA is the current real-time state; together with data from other biosensors, these will eventually be integrated into a model of "you." He expects this technology to become as common as smartphones or AI, with costs dropping exponentially—someday anyone will be able to see their DNA and RNA expression in real time.

Zooming back to the crypto world, this is exactly what "decentralized science" (DeSci) has been talking about in recent years: returning scientific research, data ownership, and genomic information from institutions to individuals. Projects like VitaDAO and BIO Protocol tokenize biotech research, funding, and intellectual property. What Bradley Woolf does is the most fundamental, most primitive step on this path: even the sequencing action itself is not handed over to medical institutions; he does it at home.

Frequently Asked Questions

Is the MinION sequencer the same as genetic testing done in hospitals?

No. The MinION is a research-grade, portable sequencing device. Bradley Woolf himself emphasizes that the first results are not yet diagnostic grade, which is completely different from hospital lab tests with clinical validation processes. They cannot replace each other.

Is it legal and safe to sequence your own genome at home?

Sequencing your own DNA is not illegal, and the equipment and reagents can be purchased publicly. However, interpreting the data requires professional knowledge. Making medication decisions or attempting gene editing based on the report involves risks that the user bears entirely on their own.