Ever since the end of the Human Genome project we have been witnessing the steady advance of genome sequencing technology and the decline of the price required to perform it. This lead to the advent of commercial personal genome sequencing services with the appearance of a few companies about a decade ago, and the creation of a market which eventually spread to a few more companies and laboratories nowadays. For under €1000 or even under €500, it is possible today for an individual to contract a direct-to-consumer whole genome sequencing service and receive this information.

However, for the informed citizen, suspicious about our current societal state of affairs regarding technology and privacy, some concerns come to mind about the guarantees of protection of a subject's genetic patrimony when pondering investing on one of these services.

By extrapolating the past trend of the cost per genome sequenced, one thought naturally comes to mind:

Is it reasonable to expect the cost of sequencing an individual's genome to drop so much in the near future as to allow for a new kind of product to be developed and sold in the coming years with which one could perform personal genome sequencing at home?

Additionally, let me also ask:

What alternative options and safety mechanisms can we implement (or foresee to be implemented) at a personal and collective level to eliminate this privacy violation concern and protect the genome data on public and private servers?

  • 1
    Like this: nanoporetech.com/products/minion ? – JMP Nov 30 '19 at 20:48
  • Well we should read the agreements we mindlessly click. You will have to block out half a day to carefully read one, and assuming you understand it, you may be able to protect yourself. But these agreements are not negotiable for small players. I don’t spend time reading agreements like this. There is no way I’m signing one regarding the subjects you ask about. – Gordon Dec 1 '19 at 14:12
  • So it is not so much a matter of data security, though that is important, but the fact that we have (probably) consented to a use we did not read about before clicking “accept”. – Gordon Dec 1 '19 at 14:16
  • @Gordon, I partly agree. But that is exactly what data protection regulation is for. With the GDPR, DNA likely falls under the special data category (special personal data) that need even more protection. Companies are only allowed to use that data for the original intended purpose and nothing more. No legal agreement can overrule that. While GDPR is not the solution to everything, it also requires the use of a short and read-/understandable privacy statements. – Bob Ortiz Dec 2 '19 at 9:26
  • @JMP Thanks for that! Do you know of any competitors? – healtech Dec 5 '19 at 15:13

I am working in the information security industry and personally following medical technology closely. Your first question kind of asks to predict the future, which nobody can. Yet, I will give you my personal opinion on the matter.

Is it reasonable to expect the cost of sequencing an individual's genome to drop so much in the near future as to allow for a new kind of product to be developed and sold in the coming years with which one could perform personal genome sequencing at home?

I would estimate that genome tech and genome research is still growing steadily and more practical uses and (medical) applications are yet to be found. Along with that, current productivity (see this graph), effectiveness and accuracy can always be improved.

That said, there is apparently a big need for this technology and as there is a need, there is inherently a need for innovation of that technology. Innovation often means increase productivity (doing more with less). The Moore's law (although currently controversial) is the observation that the number of transistors in dense integrated circuit doubles about every two years. That said, it applies directly to all technology that needs transistors (which is mostly all technology). If you want another great example of technology where effectiveness has gone up and cost has gone down drastically, check out the development of solar technology in the last 60 years (see this article). So in short, yes it is very likely that the cost will keep going down the coming decades.

What alternative options and safety mechanisms can we implement (or foresee to be implemented) at a personal and collective level to eliminate this privacy violation concern and protect the genome data on public and private servers?

Answering from an information security perspective this question can also only be answered in a very subjective way, meaning it is only an opinion. Data is data and humans are human. Sounds silly but I will explain. Ideally, you would want all sequencing devices to implement solid, future-proof and quantum-safe encryption on the data processed, stored and exported by the machine. Yet, there are two problems here.

First, data is data. Cheap tech means a consensus, when tech makers want to compete they could try to use weaker or no encryption at all. Also, how will this be enforced? By law? Then, who will enforce that law? What if people entirely make the device themself? What if the data is initially encrypted but later in the process handled or stored insecurely?

Secondly, humans are human. As with all data protection and information security, the human is the weakest link in the chain. Data might be handled correctly at first but can you expect from an average user to deal with complicated technical processes and will it still be as secure when those processes are made extremely simple and user-friendly? By law, at least in the European Union with the General Data Protection Regulation (GDPR), companies are responsible to protect your data. But, once they hand it over to you in a secure way and destroy it themself, it becomes entirely your own responsibility.

Additionally, what if (some) people want to share this information for medical reasons. Imagine the massive amounts of data, where a great number of correlations can be found between diseases or where new unidentified diseases and currently unknown causes of diseases can be found. Imagine the information you'll get from the DNA of people with genetic immunity to certain diseases.

I will exaggerate a bit now, my apology for entirely ignoring the privacy aspect here but: Would it be ethical to keep all this data private? Or is the only ethical thing to make it all part of one global scale medical research? Imagine the practical use cases of all this data. Should it be an acceptable risk? What is the actual risk of "leaking" your DNA data anyway? Using global research results for designer DNA with CRISPR? You leave your physical DNA everywhere you go (see this article). Do we clean up all our hairs, skin cells, other bodily fluids that we left in all public places to prevent people from physically finding it? No! So, why would we do otherwise digitally in the information age?

| improve this answer | |
  • This is a great answer, but it's mainly an opinion piece that lacks supporting references. Supporting references are required here, so could you please find one or two relevant to your main points? – Carey Gregory Dec 2 '19 at 5:31
  • 1
    @CareyGregory thanks for the feedback. I wrote this answer on my phone when I was waiting for something. Not the best moment to find sources and reference them. Anyway, I've added a few. Better now? – Bob Ortiz Dec 2 '19 at 9:27

The first question relates to the retail price floor of genome sequencing hardware. As with any commercial product, price depends on demand. I don't envision much demand for home genome sequencing, for three reasons.

  1. The hardest part of genomic inquiries is not getting the data, but interpreting it. The demand for interpretation will grow, and I suspect that is where the competition will be. The interpretation products will require your data in order to operate, and it will always be easier to send them a cheek swab than to put a complicated piece of machinery in your home to get the data.

  2. If medical decisions are to be made on the basis of home genomic results, then they would have to yield very high quality results. You can probably build a jet engine at home, but you wouldn't fly in an airplane powered by your engine. So I don't see demand for home genomic medical testing matching even the demand for home blood pressure testing, upon which decisions are being made today, but always with a double-check obtained in the clinic.

  3. If the intended use is non-medical, there may be market overlap with portable research systems (e.g. Nanopore). The price points in that market will be higher than for consumer home uses, if for no other reason than researchers will get more value from portable testing than consumers. You might use the machine once on yourself and on your family, friends, and pets, and never again.

The second question relates to protecting genomic information once it has left the confines of the home. It's impossible to do with certainty. If an attractive market for genomic information develops, then even trusted people and enterprises will violate that trust, despite whatever legal penalties may ensue, and despite whatever technological controls are in place. Information is information, and if it exists somewhere in readable form for any period of time, it can be copied. The only question is how many people need to be bribed or otherwise corrupted. Sysadmins are not incorruptible.

| improve this answer | |
  • Thank you for your opinions. They helped me get more convinced that: yes, it will soon be possible to sequence our whole genome with a portable device, at home, and yes, it will be possible to interpret it with personal software and hardware, at home. Challenge will be to find cross correlations across populations in non-centralized patient data. – healtech Jan 29 at 12:19
  • Also, I don't mean to start a discussion to refute your fragile arguments but let me just ask you a question: what do you think about end-to-end encryption? – healtech Jan 29 at 12:26
  • (1) It's who's at the ends that makes the difference. If you submit a sub-sequence to some company for interpretation, they are an "end" and they now have your subsequence forever, and cannot be trusted to do the right thing with it. As multigenic characteristics become better understood in the future, getting an analysis will involve uploading more and more of your sequence: it won't be SNPs. (2) You only have to slip up once, or the encryption only has to fail once, and that's it -- there is no going back. But maybe you're perfect in all you do, including picking flawless technology. – Iron Pillow Feb 7 at 16:37
  • Can't you envision analytics in personal computers? This could be done with interpretation markers prepackaged in the software, or even by sending data to the cloud through some protected channel in, I don't know, some kind of encoded format to be processed on the fly by an hard-to-explain trained neural network... Just making shit up, I'm sure these concerns are current research questions. I wouldn't be so sure as you seem to be about how it is going to play out... Are you familiar with federated learning ? – healtech Feb 7 at 18:50
  • As soon as you send the data anywhere you have lost control--it can be copied or read, even if there are pledges not to. (Great story on CryptoAG in today's Washington Post.) If you send a transform of the data, your disclosed information varies with the transform and size of data fed to the transform. If the analytical smarts come to the data, great, but (1) the "smarts" could steal it, (2) the $$ is in the smarts and so companies will be loath to implement it that way. Companies are data hungry. They want the data; as long as their analyses are best, resistance is futile. I'm certain! :-) – Iron Pillow Feb 11 at 18:01

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.