如何在家里测定自己的 DNA

如何在家里测定自己的 DNA
How to sequence your own DNA at home

原始链接: https://bradleywoolf.com/links-1/sequencing-my-own-dna-at-home

本摘要概述了使用牛津纳米孔技术（ONT）进行居家基因组测序的流程。作者通过 MinION 设备、口腔拭子及标准化的分子生物学实验室设备，已成功完成了五次个人基因组测序。 **工作流程：** 1. **准备阶段：** 包括建立无菌环境，以及采购试剂（如 DNA 提取试剂盒、文库构建预混液）和设备（如 Qubit 荧光定量仪、微型离心机）。 2. **测序阶段：** 通过拭子采集细胞，进行裂解以提取高分子量 DNA，随后进行修复并连接接头。在完成文库构建和质量控制后，将样本加载至流动槽进行实时测序。 3. **分析阶段：** 原始数据（POD5 文件）通过 *Dorado*（碱基识别）、*minimap2*（比对）和 *Clair3*（变异检测）等软件进行处理。随后将生成的 VCF 文件与 *ClinVar*、*VEP* 和 *PharmGKB* 等生物数据库进行比对查询。 **核心理念：** 基因组是“参考层”而非诊断手册。尽管目前的居家测序能力尚达不到医疗级标准，但它已能帮助个人查询自身的 DNA 信息，从而深入了解药物代谢、罕见变异和基因通路。作者强调，该领域正迅速成熟，正朝着集成式实时生物传感的未来发展。

这场 Hacker News 讨论聚焦于家庭 DNA 测序的实用性与伦理问题，起因是一篇关于使用牛津纳米孔（ONT）技术的指南。 **技术可行性：** 虽然使用 ONT MinION 等设备在物理上可以实现家庭测序，但专家警告称其成本高昂且技术难度大。一个主要顾虑是数据准确性；ONT 的错误率具有系统性而非随机性，这意味着增加“测序深度”并不能像短读长测序那样有效地消除错误。 **隐私与效用：** 许多用户寻求家庭测序是为了完全掌控自己的遗传数据。然而，该指南建议使用大语言模型（如 Claude）进行分析，这在注重隐私的硬件与基于云端的软件依赖之间造成了矛盾。 **动机：** 关于家庭测序价值的观点两极分化。支持者将其视为个性化医疗和健康干预的里程碑。批评者则将其斥为“自我陶醉”或危险行为，并指出未经培训的用户缺乏解读遗传数据的背景知识，无法安全使用。此外，人们还对获得可执行结果所需的医疗级质量控制提出了担忧，许多人认为专业实验室仍然是临床诊断的标准。

I have now sequenced my own genome 5 times with an Oxford Nanopore Technologies MinION. This means collecting them from a swab, prepping them for sequencing, running them through a sequencer, then doing analysis over them.

Cheek cells are easily accessible and replenish pretty quickly. They are not used for cancer diagnosis, inflammation, or what genes are being activated in other parts of the body (like if you have hives on your chest and want to test what genes are being expressed in the cells that are inflamed), since you would want to collect the cells having problems and compare them against other normal versions of those cells.

To sequence the cells, I bought lab materials and consumables to sequence my own genome at home. It took me about two months to get everything together to do a full end to end high quality run. Likewise, the costs are still out of reach for the average person but they are decreasing (exponentially!) and we will eventually have affordable technology, like a cell phone or AI, telling us about our DNA + RNA expression real-time.

Before we actually spend all this time and money on sequencing, what can we actually do with our genome?

The genome is not magic by itself- it is the reference layer. Once I have a VCF, I can run it through tools like VEP, ClinVar, gnomAD, PharmGKB (highly recommend), Gene Inspector, or Claude, and start asking:

Which variants do I have?
Which genes and pathways are affected?
Which medicines might I metabolize differently?
What rare variants should I take seriously?
Where does the model know nothing yet?

This last part matters- the information produced is not yet diagnosis-level, and it is definitely not “edit yourself with CRISPR because an AI said so.” The near-term value is turning a static genome into something queryable, but the “edit yourself with CRISPR” will most likely follow. DNA is the stable reference, RNA is the current state, and we will eventually integrate all biosensor data into one ‘model’ of yourself.

Here are a list of links and Twitter posts to help:

Genetic variants from first principles:

TLDR these genes, and the combinations of them, compound into real physical diseases. We will probably map these in the next decade

Pass your genome + RNA to any of these models: https://www.biotender.online/bio-model-install-guide/
Give your genome to Claude Code- message me if you want me to set this up for you
Take it to a doctor if you metabolize particular drugs differently
Patrick Collison post on using agents to talk to his genome

This is intended to be read by AI- please just copy and paste the URL of this and have ChatGPT walk you through it. If you have AR glasses, even better, since the AI can walk you through the whole protocol.

Hardware

Consumables

Reagents

NEB Monarch HMW DNA Extraction Kit for Cells & Blood ($87 for 5 runs)

Nuclei Prep Buffer
Nuclei Lysis Buffer
RNase A
Proteinase K
Precipitation Enhancer
DNA Capture Beads
gDNA Wash Buffer
Elution Buffer II
Bead retainers / Monarch collection parts from kit

DNA library-prep reagents

FFPE DNA Repair Buffer
FFPE DNA Repair Mix
Ultra II End-prep Reaction Buffer
Ultra II End-prep Enzyme Mix
NEBNext Quick T4 DNA Ligase

Oxford Nanopore SQK-LSK114 ($720 for 6 reactions)

Long Fragment Buffer / LFB
Elution Buffer / EB
Ligation Adapter / LA
Ligation Buffer / LNB
Sequencing Buffer / SB
Library Beads / LIB

Flow cell priming reagents

Flow Cell Flush / FCF
Flow Cell Tether / FCT
BSA

AMPure XP beads

80% ethanol

Nuclease-free water ($32 for 25mL)

DNA quantity measurement

Qubit fluorometer
Qubit dsDNA BR or HS Assay Kit

Bench equipment

Microcentrifuge
Vortex mixer
Heat block / dry bath
Magnetic rack for 1.5/2 mL tubes
Tube racks
Ice bucket / cold block
Freezer at -20°C
Fridge at 4°C
Pipettes

P1000 (100–1000 µL)
P200 (0–200 µL)
P20 (2–20 µL)
P10 (1-10 µL)

Plastics / Lab Pro-style consumables

Sterile flocked cheek swabs
1.5 mL microcentrifuge tubes
2.0 mL microcentrifuge tubes
DNA LoBind 1.5 mL tubes
RNA LoBind tubes
0.2 mL PCR tubes
Qubit assay tubes
Pipette sterile filtered tips

P1000 (100–1000 µL)
P200 (0–200 µL)
P20 (2–20 µL)
P10 (1-10 µL)

Wide-bore P200 tips
Tube labels / lab marker
Gloves

Software stack

MinKNOW
Dorado
minimap2
samtools
mosdepth
NanoPlot or pycoQC
Clair3
DeepVariant, optional
Ensembl VEP
ClinVar
gnomAD
PharmGKB
dbSNP, optional
Python/R
SQLite/Postgres for query layer later

The goal is to go from 2 cheek-swab samples → MinION sequencing

Gloves on.
Clean bench.
Label tubes:

Cheek sample
Extracted DNA
End-prep
Ligation
Final library
Priming mix
Loading mix

Bring AMPure XP beads to room temperature.

Keep enzyme mixes cold.

Set heat block to 56°C.

Keep Nuclei Prep Buffer cold.

Confirm gDNA Wash Buffer already has ethanol added.

Confirm you have isopropanol for DNA binding.

Confirm you have fresh 80% ethanol for AMPure cleanup.

Confirm you have the corrected ONT reagents:

FFPE DNA Repair Buffer v2
FFPE DNA Repair Mix
N-Prep Enzyme Mix
Salt-T4 DNA Ligase
LNB
LA
LFB
EB
SB
LIB
FCF
FCT

Goal: get as much cheek-cell material as possible into PBS.

Rinse mouth with water.
Wait 10 minutes.
Do not brush teeth.
Do not use mouthwash.
Scrape inside cheek firmly for 60 seconds.
Add 1 mL cold PBS to labeled tube.
Put swab head into PBS.
Vortex 10 seconds.
Press swab against tube wall to squeeze liquid out.
Remove and discard swab.

What good looks like:

PBS may look slightly cloudy.

Goal: concentrate cells and remove excess PBS.

Spin at 2,000 × g for 30 seconds.
Look for a small white/off-white pellet or smear.
Remove most PBS with P1000.
Remove more PBS carefully with P200.
Leave 50–100 µL above the pellet.
Flick gently to resuspend the pellet.

Do not aspirate the pellet.

Nuclei Prep Solution

165 µL Nuclei Prep Buffer
5.5 µL RNase A

Mix gently. Keep cold.

You will use 150 µL.

Nuclei Lysis Solution

165 µL Nuclei Lysis Buffer
11 µL Proteinase K

Mix gently. Keep at room temperature.

You will use 150 µL.

The extra volume is intentional so pipetting error does not leave you short.

Goal: break open cells and digest proteins while preserving long genomic DNA.

Add 150 µL Nuclei Prep Solution to the resuspended cheek-cell pellet.
Pipette up/down gently 10 times.
Incubate 2 minutes at room temperature.
Add 150 µL Nuclei Lysis Solution.
Invert tube gently 10 times.
Do not vortex.
Incubate at 56°C for 10 minutes.

What good looks like:

The liquid may become more viscous.

Do not vortex after lysis. At this point, the priority is preserving high molecular weight DNA.

Goal: precipitate genomic DNA onto the large Monarch capture beads.

Add 75 µL Precipitation Enhancer.
Invert gently 8–10 times.
Add 2 DNA Capture Beads.
Add 275 µL isopropanol.
Invert slowly 30 times.
Do not vortex.

Important:

The beads now carry the DNA.
Do not lose the beads.
Do not use ethanol here.

Goal: wash contaminants away while keeping DNA bound to the capture beads.

Let beads settle briefly.
Remove liquid carefully without removing beads.
Add 500 µL gDNA Wash Buffer.
Invert gently 2–3 times.
Remove wash buffer carefully.
Add another 500 µL gDNA Wash Buffer.
Invert gently 2–3 times.
Remove wash buffer carefully.
Remove as much residual wash as practical without touching or removing beads.

Critical:

gDNA Wash Buffer must already have ethanol added.

Goal: release purified genomic DNA from the Monarch capture beads.

Put bead retainer into Monarch collection tube.
Transfer/pour beads into the retainer.
Pulse spin ≤1 second.
Move beads to a clean Monarch tube.
Add 100 µL Elution Buffer II.
Incubate at 56°C for 5 minutes.
Put bead retainer over a clean DNA LoBind tube labeled Extracted DNA.
Transfer eluate + beads into bead retainer.
Spin at 12,000 × g for 30 seconds.
Keep the eluate.

The eluate is the purified genomic DNA.

Do not discard the eluate.

Goal: measure whether there is enough DNA to justify ONT library prep.

Use:

1x dsDNA High Sensitivity

Do not select the older non-1X dsDNA HS workflow if using the premixed 1X reagent.

Standards

Standard #1:

190 µL 1X dsDNA HS working solution
10 µL Standard #1

Standard #2:

190 µL 1X dsDNA HS working solution
10 µL Standard #2

Sample, first attempt

198 µL 1X dsDNA HS working solution
2 µL DNA

On Qubit, enter:

Sample volume = 2 µL

If too low

Use more DNA in a fresh Qubit tube:

190 µL 1X dsDNA HS working solution
10 µL DNA

On Qubit, enter:

Sample volume = 10 µL

Do not just add 8 µL to the old 198 + 2 tube, because that makes 208 µL total and breaks the assay math.

Record:

Genomic DNA concentration:
Remaining DNA volume:
Estimated total DNA:

Calculate:

total DNA ng = Qubit concentration ng/µL × remaining volume µL

This is the best place to pause. For a same-day break:

Store extracted DNA at 4°C / fridge.

Before storing:

Make sure the DNA is in a clearly labeled DNA LoBind tube.
Quick spin.
Do not vortex.
Close tube tightly.
Put it in the fridge.

Resume at Step 10.

Goal: prepare 47 µL DNA input.

Ideal target:

1,000 ng DNA in 47 µL

Calculate:

DNA volume needed = 1000 ng / Qubit concentration ng/µL

Then:

water volume = 47 µL - DNA volume

If the DNA is too dilute and 1,000 ng cannot fit into 47 µL, use the maximum possible volume:

47 µL extracted DNA
0 µL water

For the first low-input run:

0.296 ng/µL × 47 µL = ~13.9 ng DNA

That was far below the recommended input, but useful as an end-to-end practice run.

Goal: repair DNA and prepare the ends for adapter ligation.

Use the actual v2 reagents:

FFPE DNA Repair Buffer v2
FFPE DNA Repair Mix
N-Prep Enzyme Mix / Ultra II End Prep Enzyme Mix

Do not use Salt-T4 DNA Ligase in this step. It is used later.

If not using DCS, replace the optional 1 µL DCS with 1 µL nuclease-free water.

Reaction

47 µL DNA
1 µL nuclease-free water
7 µL FFPE DNA Repair Buffer v2
2 µL FFPE DNA Repair Mix
3 µL N-Prep Enzyme Mix

Total = 60 µL

Steps

Add 47 µL DNA to a 0.2 mL thin-walled PCR tube labeled End-prep.
Add 1 µL nuclease-free water.
Add 7 µL FFPE DNA Repair Buffer v2.
Pipette mix 10–20 times.
Add 2 µL FFPE DNA Repair Mix.
Pipette mix 10–20 times.
Add 3 µL N-Prep Enzyme Mix.
Pipette mix 10–20 times.
Quick spin.

Incubate:

20°C for 5 minutes
65°C for 5 minutes
Then place on ice

Goal: clean the repaired/end-prepped DNA.

Input:

60 µL repair/end-prep reaction

Resuspend AMPure XP beads until uniformly brown.
Add 60 µL AMPure XP beads to the 60 µL reaction.
Mix gently by pipetting 10 times.
Incubate 5 minutes at room temperature.
Put tube on magnet.
Wait until solution clears.
Remove and discard supernatant without touching beads.

Wash 1

Keep tube on magnet.
Add 200 µL fresh 80% ethanol.
Remove ethanol.

Wash 2

Add another 200 µL fresh 80% ethanol.
Remove ethanol.

Dry

Remove residual ethanol with P10/P20.
Air dry briefly, about 30 seconds.
Do not overdry.
Do not let beads crack.

Elute

Remove tube from magnet.
Add 61 µL nuclease-free water.
Resuspend beads gently.
Incubate 5–10 minutes at room temperature.
Put tube back on magnet.
Wait until clear.
Transfer 60 µL clear eluate to a clean DNA LoBind tube labeled Ligation.

Do not transfer beads.

Goal: attach ONT sequencing adapters.

Use the actual reagents:

LNB
LA
Salt-T4 DNA Ligase

Reaction

60 µL repaired/end-prepped DNA
25 µL LNB
10 µL Salt-T4 DNA Ligase
5 µL LA

Total = 100 µL

Steps

Use a clean DNA LoBind tube labeled Ligation.
Mix LNB slowly by pipetting before adding. It is viscous.
Add 60 µL repaired/end-prepped DNA.
Add 25 µL LNB.
Add 10 µL Salt-T4 DNA Ligase.
Add 5 µL LA.
Mix gently by pipetting 10–15 times.
Do not vortex.
Quick spin.
Incubate 10 minutes at room temperature.

Before incubation, say/check:

DNA, LNB, Salt-T4 DNA Ligase, LA.

Failure points:

Forgetting LA = no sequenceable library.
Forgetting Salt-T4 DNA Ligase = adapter ligation fails.
Poorly mixed LNB = bad ligation chemistry.
Vortexing = unnecessary DNA shearing.

Goal: remove free adapters, ligase, salts, and small fragments while keeping adapter-ligated DNA.

Important:

This cleanup uses LFB, not ethanol.

Input:

100 µL ligation reaction

Resuspend AMPure XP beads fully.
Add 40 µL AMPure XP beads to the 100 µL ligation reaction.
Mix gently by pipetting 10 times.
Incubate 5 minutes at room temperature.
Place tube on magnet.
Wait until solution clears.
Keep tube on magnet.
Remove and discard supernatant carefully.
Do not touch bead pellet.

Wash 1

Add 250 µL LFB to beads.
Remove tube from magnet.
Gently resuspend beads by flicking or slow pipetting.
Put tube back on magnet.
Wait until clear.
Remove and discard LFB.

Wash 2

Add another 250 µL LFB.
Remove tube from magnet.
Gently resuspend beads.
Put tube back on magnet.
Wait until clear.
Remove and discard LFB.
Remove residual LFB with P10/P20.
Do not overdry.

Elute

Add 25 µL EB.
Remove tube from magnet.
Resuspend gently.
Incubate 10 minutes at room temperature.
Put tube back on magnet.
Wait until eluate is clear.
Transfer clear eluate to a clean DNA LoBind tube labeled Final library.

This is the final sequencing library.

Practical rule:

Liquid moves. Beads stay.

Do not transfer beads into the final library.

Goal: measure final adapted library concentration.

Because low-input runs may produce very little final library, this may read low or fail.

Use:

199 µL 1X dsDNA HS working solution
1 µL final library

On Qubit:

Sample volume = 1 µL

Record:

Final library concentration:
Final library volume:
Estimated mass loaded:

Calculate:

mass loaded ng = final library concentration ng/µL × 12 µL

If the final library is too low, do not repeatedly burn more library on Qubit. For a practice run, proceed with 12 µL library in the loading mix.

Goal: verify the flow cell before loading the library.

Take flow cell out of fridge.
Let it sit at room temperature for 20 minutes.
Keep it flat.
Do not shake.
Plug in MinION.
Open MinKNOW.
Insert flow cell.
Run flow cell check.
Record active pores.

Decision table:

>1200 pores = great
800–1200 pores = usable
500–800 pores = marginal/practice
<500 pores = bad, but still possible for mechanical practice
<200 pores = essentially not worth it except loading practice

Record:

Flow cell ID:
Starting active pores:
Flow cell age:
Previously used? yes/no:
Washed? yes/no:

At the end, there are three tubes:

Tube 1: Final library
- This is the DNA library made after adapter cleanup.

Tube 2: Priming mix
- This prepares the flow cell.

Tube 3: Loading mix
- This contains final library + sequencing buffer + library beads.

There are two flow-cell ports:

Port 1: Priming port
- This is under the sliding cover.
- Priming mix goes here.

Port 2: SpotON sample port
- This is the small sample well.
- Loading mix goes here drop by drop.

Important:

Final library does not go directly onto the flow cell by itself.
Final library first goes into the loading mix tube.

If BSA is available:

1170 µL FCF
5 µL BSA
30 µL FCT

Total = 1205 µL

If BSA is not available and this is a practice run:

1170 µL FCF
30 µL FCT

Total = 1200 µL

Do not substitute SFB, DCS, LIS, or LNB for BSA.

Mix gently. Avoid bubbles.

This is Tube 2: Priming mix.

Do this only after flow cell check is complete. Make sure to watch this video: https://www.youtube.com/watch?v=IknVaEnuDz0. It is how you load the actual flow cell!

Keep flow cell plugged in and flat.
Open the sliding cover to expose the priming port.
Check for an air bubble near the port.

Draw back only 20–30 µL:

Set P1000 to 200 µL.
Put tip into priming port.
Slowly dial to 220–230 µL.
Stop as soon as liquid enters the tip.
Do not pull more.

Then load into the priming port:

Go slowly. Avoid bubbles.

Wait:

Tube 3: Loading mix

Use:

37.5 µL SB
25.5 µL LIB
12 µL final library

Total = 75 µL

Important:

LIB = Library Beads from the ONT kit.
LIB is not AMPure XP beads.
LIB settles fast.
Mix LIB immediately before pipetting.

Steps:

Mix LIB immediately before pipetting.
Add 37.5 µL SB to a clean tube.
Add 25.5 µL LIB.
Add 12 µL final library.
Mix gently by pipetting.
Do not vortex.

After the 5-minute wait, load into the priming port:

Avoid bubbles.

Gently mix Tube 3 loading mix immediately before loading.
Open the SpotON sample port.
Load the full 75 µL loading mix into the SpotON port.
Add it drop by drop.
Let each drop disappear before adding the next.
Do not force it.
Do not jab the port.
Avoid bubbles.

Then:

Close SpotON cover.
Close priming port.
Add light shield.
Close MinION lid.
Start run in MinKNOW.

Final mental model:

Priming mix → sliding priming port
Loading mix → SpotON sample port
Final library → only goes into loading mix first

Recommended basic settings:

Flow cell type: FLO-MIN114
Kit: SQK-LSK114
Basecalling: ON
Model: High-accuracy / HAC
Barcoding: OFF
Alignment: OFF
Adaptive sampling: OFF
Raw reads / POD5: ON
Filtering: OFF for low-input practice runs

Then:

Save configuration
Start

For a low-input practice run, keep raw POD5 on so the data can be analyzed later even if live output is poor.

Find POD5 output directory.
Install Dorado.
Run basecalling:

dorado basecaller sup pod5_directory/ > calls.bam

Convert BAM to FASTQ if needed:

samtools fastq calls.bam > reads.fastq

For faster first pass, use HAC instead of SUP:

dorado basecaller hac pod5_directory/ > calls.bam

Download GRCh38 FASTA.
Index reference:

minimap2-d GRCh38.mmi GRCh38.fa

Align reads:

minimap2-ax map-ont GRCh38.mmi reads.fastq | samtoolssort-o aligned.bam

Index BAM:

samtools index aligned.bam

Check alignment summary:

samtools flagstat aligned.bam > flagstat.txt

Check coverage:

mosdepth sample_cov aligned.bam

Install Clair3.
Use ONT model.
Run Clair3 with GRCh38 reference, sorted BAM, and output directory.
Output should include VCF.
Do not overinterpret low-coverage variants.
For a first MinION run, treat this as technical validation, not medical-grade interpretation.