Nick Tsimplis

23andMe currently have a special Christmas sale on their flagship “Health + Ancestry” DNA sequencing service, offering it for only £99. I can’t say I understand the appeal of receiving a DNA test from a relative. It has a certain feeling of “Aha, I knew we weren’t related”. But with 23andMe reporting to have over 10 million customers so far I’m sure my non-purchase won’t bother them. I can however understand the appeal of having access to the genetic data of 10 million individuals given the ever increasing value of data. So what does 23andMe know and what do they do with it?

Lets Start with What They Know

23andMe and Ancestry.com do not offer full genome sequencing and they don’t check every single part of your DNA. Instead they use single nucleotide polymorphism (SNP) testing which works by checking if certain known modifications have occurred to important regions of the DNA. These changes are usually inherited from parents and can be indicators of ethnic origin. They can also be indicators of predisposition to disease. Companies that test the entire genome do exist but are much more expensive and often need a doctor’s referral.

The SNP method is chosen because it specifically targets regions of interest. As a result it provides maximum results for minimal work and cost and allows the sequencer to acquire a large amount of genetic data about a number of different factors. Carrier status of genetic diseases, health risks, genetic traits, ethnic origin and overall wellness can be, to a varying levels of accuracy, be predicted with this information.

How Accurate Is This Information?

The SNP method has a severe limitation in that it cannot identify unexpected changes in the genome. It can thought of like searching a word document for spelling mistakes. If you search for a typo you will only find instances of that exact typo. So you will find the word “color” because it’s a common typo and you know to look for it. But you won’t find “collour” unless you specifically search for it or manually check every word. It’s the same with SNP genome testing. It will only tell you it has not found a genetic variation not that you don’t have one. This makes the test very unreliable in cases where an unexpected mutations occur.

There is also the influence of environmental factors. Whilst a gene variant may increase chance of heart disease, your diet and lifestyle are equally significant. Because of this the results are often not given as definite predictions but rather as an increased likelihood of having or developing traits. In the case of some traits, like caffeine consumption, it may be a minor issue but when it comes to health factors the consequences are more significant.

In 2013, the FDA acted on these concerns. They issued a letter to 23andMe ordering them to stop marketing their products as health predictors. The FDA felt that the tests were not accurate and that customers were not properly informed about what the results meant. In response, 23andMe made changes and in 2015 received approval to test for Bloom syndrome, a genetic disorder caused a single change in the patient’s DNA. 23andMe is currently approved to test for 10 genetic diseases. However they are still restricted from informing customers about predisposition to cancer and other multi-factor diseases.

So What Do They Do My Genetic Data?

In short, they sell it. Much like many big tech companies 23andMe’s business model is primarily focused on data collection and distribution. It has made the majority of it’s money from selling the data of it’s users. It’s able to do this due to a line in the terms and conditions that allows individuals to “donate” their genetic data for use in research.

23andMe already has deals with several pharmaceutical companies with one of the first notable ones being with Genentech in 2015. Access to 23andMe’s database of 800,000 customers cost Genentech a total of $60 million. In March 2018 a larger deal with pharmaceutical giant GlaxoSmithKline (GSK) was announced. GSK’s $300 million investment was enough to provide them with access to the genetic data of 80% of 23andMe’s customers. With genetic therapies increasingly looking like the future of therapy, GSK’s interest is unsurprising. The GSK deal did however clearly illustrate the value of 23andMe’s data and was the primary factor in 23andMe being valued at $2.5 billion in May 2018.

Other companies have allowed access to their records to government agencies. In January 2019 Buzzfeed News reported on genetic testing company Family Tree DNA sharing their database with the FBI in violent crime cases . Whilst, Family Tree had a month earlier changed their terms and conditions to allow for law enforcement access to their data it is likely that co-operation between companies and law enforcement had begun much earlier.

In 2017 another genome database, GEDmatch, had been used to identify the Golden State Killer. By comparing a sample of the suspect’s DNA, with the database law enforcement agents were able to identify Joseph James DeAngelo. Whilst DeAngelo’s DNA had not been uploaded to the database it was a close match to one of his relatives who had uploaded their own. By identifying the suspect’s family it was then simply a case of finding a member of the family in the area the murders occurred. At this time DeAngelo’s trial has not yet taken place.

So Who Owns My Genetic Data?

Well it’s complicated. When it comes to the genetic data held in DNA it belongs to no-one. In June 2013 the Supreme Court ruled that human genes are a product of nature and as a result cannot be patented. But once it’s sequenced it becomes data that can be treated as any other bit of sensitive information, like age or ethnicity.

Because of this some legislative protection does exist. In the EU this takes the form of the General Data Protection Regulation (GDPR), that covers genetic data. But in the US no such legislation exists. Instead, the Genetic Information Non-Discrimination Act offers some protection from “genetic discrimination in health insurance and employment” and makes it illegal for health insurers to make any decision based on genetic tests. However, that doesn’t mean that you own the information in your genes. That was made clear in the case of John Moore vs University of California, Los Angeles.

In 1976, John Moore had his spleen removed as part of his treatment for leukemia. The spleen was due to be destroyed but Moore’s doctor David Golde realised that it produced a protein that had potential as an anti-cancer drug. Golde acquired a patent on the cells that produced the protein and made an agreement with the Genetics Institute which included Golde receiving 75,00 shares. The Genetics Institute began to develop the drug at which point Moore realised what was going on and sued. Eventually the case reached the Supreme Court which ruled that Moore had no right of ownership over his discarded cells. This was in 1990 by which point Golde’s shares were worth $ 3 million. In 2001 Moore passed away from his leukemia. He only ever received a token settlement from UCLA for using his cells without his consent.

So Should I Pay £99 To Get My DNA Sequenced?

Probably not. There are certainly very good reasons to have your genome sequenced but they are usually accompanied with a recommendation from a doctor. Genetic counselling is also an option if you’re unsure about if you should have your genome sequenced. However, if you’re interested the first step would always be to know what you sign up for. Read the terms and conditions very carefully and make sure you’re not allowing the testing company to share your information.

It really is best to not think of 23andMe as a gene sequencing company but as the newest member of the big tech company club. Compare them to Facebook and the parallels are clear. Whereas Facebook offers you the opportunity to tell the world about you. 23andMe offers you the opportunity learn more you. Facebook distributes your data to entice advertisers and drive revenue. 23andMe sells your data to pharmaceutical companies to drive revenue. At their core they have the same business model. To collect data and sell data.