Paternity Accuracy from 23andMe


Home Topic overview / Sitemap Webmaster



DNA testing and privacy

Author: Philipp Schaumann - Last changes: Feb. 2020

Different types of DNA tests (DNA analysis, genome analysis)

When discussing the interplay between DNA analyzes and privacy, it was important until mid-2018 to clearly separate the 2 main types.

The first type of DNA test concerns the Identity verification where, for example, the investigation of a crime involves reliably assigning DNA traces from the crime scene (flakes of skin, sperm, saliva, blood or other body fluids) to a person (more or less). Non-coding DNA is used for this, so-called STRs (short tandem repeats), parts of so-called junk DNA.

The databases of (more or less) suspects created by the police authorities in many countries contain information about these STRs, i.e. they did not allow any statements until mid-2018 but allowed health, hair color or ethnic classifications - the goal was always 1 to 1 until around 2018 Comparing a suspect's DNA with DNA at the crime scene.

They are closely related to this Parentage reports (paternity tests). These tests also do not collect or evaluate any information about the person's genetic makeup, but there are a number of privacy issues (see below).

The second major class of DNA analyzes are used for genealogical purposes or health analysis carried out. They are offered by a large number of (mainly US) companies, such as 23andme,, AncestryDNA, GEDmatch or Coding DNA is used for this, so-called single nucleotide polymorphism (SNP). This technique is used to find possible clues about susceptibility to disease and information about the possible origins of ancestors.

Since the end of 2015, the two processes have started to mix. In 2015 there were the first links between the genome analysis services and the identification of the police. The US authorities also started comparing crime scene DNA with the DNA databases of and 23andMe.

Link conventions:
Bold links open another page in a new window
Links with a blue background remain on

The spectacular breakthrough came in 2018: the US police entered the DNA from the crime scene of a very old unsolved case into a genealogical database and found distant relatives of the perpetrator, and from there also the perpetrator.

The NYT and other media then discussed whether this use of the genetic data for purposes other than those for which it was collected was legally and / or ethically OK. But the development continues, because Wired writes in 2018:

    . . . shown that you could link up records between the two kinds of databases, by mapping nearby SNPs to the non-coding repeats. . . . the bigger deal is that Rosenberg’s work reveals that there’s much more information contained in a forensic DNA profile than previously thought.

This means that the huge DNA databases that the authorities in many countries have set up that should not contain any genetic information obviously do so. This is very bad for privacy and shows how data, once collected, can suddenly create entirely new privacy threats due to advances in science.


1. DNA tests to establish identity

First a brief introduction to DNA:

Only very few sections of human DNA (a maximum of 5 percent) contain the blueprint for the proteins (i.e. they "code" the proteins). Other parts code for different forms of RNA, which can also be assigned to the genetic material. Most of the DNA used to be called junk DNA, today it is called non-coding DNA.

In the areas of the genome (the 23 chromosome pairs) that are not part of the genetic make-up there are extensive DNA segments whose function is unknown (or which have no function). Here you will find small, repetitive sections, the mini satellites (VNTR, variable number tandem repeats) or STRs (short tandem repeats). These DNA segments (locus) are repetitions of a specific sequence (repeats). The number of repetitions is variable. This number is examined in the genetic fingerprint. Depending on the number of repetitions, the duplicated section has a certain length, which can be represented as a single band using gel electrophoresis in the agarose gel. The repetitive portion of the VNTRs is longer (10 to 150 base pairs) than that of the STR (2 to 7 base pairs). Most of the time, 15 such DNA sections are evaluated today, plus a gender determination (the requirements for how many matches are necessary are regulated differently by law in each country).

The probability that 2 unrelated people have the same number of repetitions in several such DNA sections is low and decreases with the number of sections examined. The number of repetitions at the various points is converted into a number and can be compared with other DNA samples or saved in a gene database. This information is stored in police databases.

As of Oct 2007, the UK has already had one DNA database which covers 7% of the population, the US has 1.5% of the population covered. In almost all countries, the authorities are now trying to create the most comprehensive DNA databases possible for their citizens under the pretext of fighting terrorists or fighting crime. One of the problems here is that there is a very high level of trust in this technique of personal identification (also due to relevant television broadcasts from the field of criminal forensics). And that's only half justified.

This trust in the reliability of the tests is even less justified since from 2019 DNA analyzers are so small and affordable that individual police stations are equipped with them: Coming Soon to a Police Station Near You: The DNA DNA Magic Box ’. The creation of a DNA profile and the comparison with the police databases can then be done within minutes. The result is that "suspects" are persuaded to voluntarily give in a saliva sample to see whether their DNA was found at the scene of a break-in (or something similar). This means that the investigation does not begin with the crime, but with a person, for whom a suitable crime is then sought. It should be noted that this is not an exact science, i.e. the device does not say: "Identity 100% proven", but rather similar DNA samples are found.

As with all such identity determinations using biometrics (including fingerprints), so-called "false positives" and "false negatives" must be considered separately. "False positives" mean that someone is "recognized" even though it is not their DNA or fingerprint at all (test result positive). "False negatives" is about the test saying it isn't that person (test result negative), but it was that person.

"False negatives" are extremely rare in DNA, i.e. this procedure is extremely well suited to proving the innocence of a defendant. In the USA this is currently (2007) leading to a wave of reopened proceedings with many hundreds of acquittals, also from people waiting to be executed.

Conversely, the procedure is very ill-suited as evidence of a person's guilt. Traces of DNA can very easily be found at the crime scene without the person being there. "Contamination" with incorrect DNA can very easily happen. One source is, for example, contamination in the laboratory. The fact that the DNA of a specific person was found in a department store or apartment does not conclusively prove that this person was in that apartment. The flakes of skin may also have been carried on someone else's clothing. Therefore, this form can only be used 100% safely for proof of innocence.

The case of the phantom murderer, who was caused by the contamination of cotton swabs, proves that contamination can even have arisen in the laboratory. False tracks can also be laid very easily in this way, e.g. by leaving someone else's hair, someone else's cigarette residue or someone else's used chewing gum at the crime scene. But there are other subjectivities in the test, see the box on the left.

Often the impression is given as if the determination of identity by means of DNA is a yes / no decision. But that is not the case at all, as this graphic from the article in the New Scientist shows: Fallible DNA evidence can mean prison or freedom. Above you can see the findings, below then 4 possible perpetrators. Which of the profiles fits best with the DNA from the crime scene is very subjective. If one of the people has the "wrong" skin color or religion, the proof of innocence can become narrow.

2007 from NEC: a portable DNA analyzer from NEC. Impressively small, but I'm not sure how to create a sufficiently sterile environment at the crime scene to prevent contamination from other people's DNA.

1.1.2008: In Austria, DNA mass tests without specific suspicion have become possible due to an amendment to the Code of Criminal Procedure. In one case in D., all men from the vicinity of a place were asked to "voluntarily" submit a DNA sample. Who refuses ......

In this amendment to the Code of Criminal Procedure in Austria, the possibility of compulsory "physical examinations" of "individual groups of people" is introduced. That could be all men in a certain region or all men in Austria. Physical examinations are not limited to oral swab DNA tests; they also include blood tests and other future tests, provided that it can be assumed that the perpetrator is among these people and that this conduct will advance the investigation.

In its current form, the information collected, e.g. the DNA profile, must be destroyed after the investigation has been completed. However, for reasons of economy (because the collection is very expensive), this could of course be changed to unlimited storage at any time. It has seldom happened that an authority, once it has data on its citizens, voluntarily destroys it. In Germany, a genetic fingerprint may only be taken if a court order.

March 2009:
2 articles in the standard on the DNA problem: The perpetrator on the DNA trail and problem of evidence after KaDeWe burglary of twins.

If such DNA profiles are stored in DNA databases, no information about the state of health or the genetic material of a person can be obtained in this way, since only so-called non-coding sequences are used for these tests and no genes, the blueprints for proteins (these would only differ in the rarest of cases). However, data protectionists defend themselves against such a total coverage of the population, as in their opinion this amounts to a general suspicion against the entire population and people easily come under serious suspicion due to "false positive" errors (e.g. through deliberately laid traces) can.

August 2009:
A laboratory in Israel has shown in a study how easy it is today to lay false DNA traces. You use 2 techniques for this. If you have a little DNA from the "wrong" culprit, you can amplify it using PCR. How a layperson can do this is described below.

The really new thing about this report is the fabrication of a false DNA sample without access to traces of the person's DNA. The key to the solution is the huge DNA databases that are now being built everywhere (and which, if governments have their say, should contain the DNA of all citizens). The researchers have succeeded in "tinkering" with any DNA sample. DNA databases do not contain DNA, but the result of electrophoresis, i.e. a pattern. The researchers take blood from any person, remove the white blood cells from it and all DNA from it, and then use a library of 425 DNA snippets to create a concoction that gives the desired profile. This is then introduced into the red blood cells of the blood sample and results in the same pattern as a real DNA sample of the person who, for example, is to be blamed for a crime.

You can discover this forgery through a test kit developed by the laboratory. The artificial DNA lacks the methyl caps. The problem is that DNA tests today unfortunately have an extremely good reputation in jurisprudence and are unfortunately not only used for proof of innocence, but also for proof of guilt.

August 2010: DNA traces not as evidence, but in the investigation
The NY Times reported controversy on July 24th. The point is that traces of DNA are used not only as evidence but also as an investigative tool, at least in the UK and in some states in the US. For this, the DNA traces are not examined with regard to "matches" ("Are the DNA traces of this suspect person?"), But rather "similar" DNA profiles are searched for in a DNA database. The UK database is one of the largest in the world and contains not only the DNA of legally convicted people, but many more.

These people are then not considered suspect themselves, but there is a suspicion that they are related to the perpetrator. In the UK, for example, the list of 200 people who may be related to the perpetrator is handed over to the police and they knock on those people's doors and ask questions. "Guilt by genetic association" is what some people call it.

In the USA, the DNA profiles of convicted persons are stored in the DNA databases, the "colored" part of the population is over-represented. Only in 1 of the states is the long list not handed over to the police, but the forensic experts independently try to further shorten the list that goes to the police with the help of other information such as residence databases.

Area-wide surveys

2016: Many governments develop desires for the DNA of all of their citizens. Collecting DNA is quite simple in individual cases, a used handkerchief or cutlery or glass is sufficient. But that does not allow comprehensive coverage. This requires laws and these are passed here and there, e.g. Kuwait, where soon all tourists will also have to surrender their DNA. Update 2016: After protests, the government eased the law somewhat: only suspected criminals will need to give their DNA. But other countries also have similar desires.

Incidentally, the DNA databases of the police authorities are not suitable for providing information about the state of health of people or their ethnic origins. For this, other gene segments have to be analyzed and that still costs a lot of money, especially for mass analyzes. But that will definitely be different in 10-30 years. More on this below.


How objective are DNA reports?

June 2016:
The Atlantic brings a long, detailed article on the problem of the rather high unreliability and lack of transparency of DNA tests to prove the guilt of a defendant: The False Promise of DNA Testing. Likewise Atlantic: The Dark Side of DNA Databases. The alleged science makes defense difficult.

Another Article Should Defendants Get Access to TrueAllele's Source Code? poses the problem that an unknown algorithm makes the guilt decision. There is no way for the accused to question this decision of the algorithm.

The New Scientist had already carried out a study on the reliability of DNA evidence in 2010. The background to this is that the supposedly objective DNA evidence is increasingly being used as evidence of a suspect's guilt. New Scientist has sent a DNA sample (from a real crime) with the traces of 2 people (which unfortunately is quite common) to 17 experts who prepare expert opinions for the courts. In reality, the alleged perpetrator was convicted on the basis of these traces.

    We took a mixed sample of DNA evidence from an actual crime scene- a gang rape committed in Georgia, US- which helped to convict a man called Kerry Robinson, who is currently in prison. We presented it, and Robinson's DNA profile, to 17 experienced analysts working in the same accredited government lab in the US, without any contextual information that might bias their judgment.
    In the original case, two analysts from the Georgia Bureau of Investigation concluded that Robinson "could not be excluded" from the crime scene sample, based on his DNA profile. (A second man convicted of the same crime also testified that Robinson was an assailant, in return for a lesser jail term.) Each of our 17 analysts independently examined the profiles from the DNA mixture, the victim's profile and those of two other suspects and was asked to judge whether the suspects' profiles could be "excluded", "cannot be excluded" or whether the results were "inconclusive".
    If DNA analysis were totally objective, then all 17 analysts should reach the same conclusion. However, we found that just one agreed with the original judgment that Robinson "cannot be excluded".Four analysts said the evidence was inconclusive and 12 said he could be excluded.

The second part of the series is here: How DNA evidence creates victims of chance.

That means we have here again a case where a method, because it sounds very scientific, does not get a really justified credibility.

March 2012:
The Economist reports a test in which an identity verification was retrospectively repeated by technicians who had no prior information about the case. Their conclusions were quite different. Quote:

    . . . but the two DNA examiners in the original case both found that they could not exclude one of the three from having been involved, based on an analysis of swabs taken from the victim. . . .
    However, of the 17 examiners Dr Dror and Dr Hampikian approached — who, unlike the original two, knew nothing about the context of the crime — only one thought that the same suspect could not be excluded. Twelve others excluded him, and four abstained.

The whole thing shows that these DNA analyzes can come to quite arbitrary results and tend to confirm the prejudices of the investigators.

Oct. 2015:
More and more doubts about the infallibility of DNA tests: 'Great Pause' Among Prosecutors As DNA Proves Fallible and Your Relative’s DNA Could Turn You Into a Suspect. And The Surprisingly Imperfect Science of DNA Testing, Subjectivity and bias in forensic DNA mixture interpretation.





2. DNA tests for parentage reports (paternity test)

This section covers paternity tests that are performed for the purpose of determining paternity only. It is now much more common that surprises with regard to paternity arise more or less randomly in the context of genealogical DNA tests, which will be discussed further below.

The doubt as to whether he is really the biological father of his children has preoccupied men for many millennia. The first test options were available after the blood group was discovered, but these tests can only be used for exclusion as a father. Today, this question can be answered quite precisely using DNA tests based on the same procedures as for establishing identity. This means that these methods are also based on so-called junk DNA, which typically does not allow any conclusions to be drawn about diseases and the like.

Uncertainties regarding the informative value arise, for example, in situations in which a twin brother could be involved. The safest statements are possible if the DNA of the child / fetus is examined as well as that of the mother and possibly other fathers. Due to the mixing of the genes of father and mother, the comparison with the child's DNA results in probability statements that are more reliable the more gene locations (individual sections of junk DNA) are included. For the matches or disagreements found, an expert then calculates the probability that these mutations are coincidental or that they are actually based on biological parenthood.

December 2010:
New Scientist points out another problem with such tests: The danger of unreliable paternity tests. If a woman wants to have such a test while there is still time for a possible abortion, a DNA sample must usually be taken from the fetus. However, this is associated with a (low) risk for the fetus. For this reason, a test has been offered since 2000 in which only the blood of the mother and possible fathers is examined. The mother's blood always contains some cells from the fetus. The problem lies in separating these from the mother's much more common cells. In the case of a male fetus this is easier, in the case of a female fetus it is very difficult and often does not succeed. The result is often false statements. New Scientist has tested such a service and has had very poor results. There are now numerous cases where mothers have married the alleged father on the basis of incorrect statements about the father, but after the birth it has clearly been shown that this is not the biological father of the child - with unpleasant consequences for the family situation.

Another aspect: Also in New Scientist in October it was reported that a man had been suspected of rape despite having a good alibi (he was in prison at the time in question). The solution: His DNA matched the real culprit's DNA because the real culprit gave the other man a bone marrow donation.


Secret DNA Testing and the Law

In D. clandestinely conducted DNA paternity tests are illegal, but at the same time it was confirmed that the legal father must be able to have his biological parenthood checked.

April 2009:
The Genetic Diagnostics Act was passed in D. It stipulates that a secret parentage test is punished with up to 5000 euros, the consent of mother and child must be present. However, if this consent is refused, there are legal enforcement options. In Austria, these tests are a gray area, but they are offered.


Illegal options allow clandestine testing and are deeply private

Secret tests, also possible in Germany

2020: Of course, secret DNA tests are not only banned in England, but also in other (European) countries. In Germany, a team from carried out a test on this: DNA test in self-experiment: Secret paternity test. You write: A DNA test on the genealogy platforms and MyHeritage now costs only 59 to 79 euros. The testers sent in the DNA of one person under 2 names, the two were classified as twins. Getting to someone else's DNA is very easy within a family.

The testers were also scared of what the providers would do with the DNA data. The business is supposedly going down (everyone who is interested in something may have already done a test). Now the companies are laying off employees and looking for other business areas for their data.

The UK Human Tissue Act states that a person's DNA can only be tested if the person [or a legal guardian (the dubious father?) Agrees]. The background to this law was, among other things, the fear of celebrities that reporters would be constantly on the lookout for DNA traces.

Probably because of the unclear legal situation (in the USA and in other countries), the secret DNA analysis for infidelity or paternity has been booming in the USA since 2008. Such services specifically advertise private customers.

A report in NewScientist Who's testing your DNA? about the situation in England is very illuminating. Such tests are banned there, but since this problem is largely unregulated in the USA, a business model around secret DNA tests has developed there: e.g. Test for infidelity in your partner, "Seeking the trues where it matters", Discreet Paternity Testing. Some of these services even have offices in London, so English customers don't even have to send the DNA samples abroad.

Another article then sheds light on the real situation in England (despite the ban): Could your DNA betray you ?. The reporter pretended to be a husband who has found a stain on the sheet and wants to have it tested to see whether this stain is really his. He stated that he also had a wine glass that his friend had used, to see if the DNA on that wine glass could be compared to the DNA of the stain.

Under the idea of ​​English law, both tests are forbidden. In the second test with the wine glass, the majority of companies advertising in England also had problems, in the first test fewer. For example, there was the argument that the inquirer only wants to know whether the DNA is in the spot of him. However, this test nullifies the idea of ​​the law. And companies that run an English website and a local mailbox, but where the actual company headquarters are in the USA and which carry out the tests there, had no qualms about it. In New York, such tests are banned only for New York citizens.

This shows that at least without internationally uniform laws, which are also executed, the bottle devil can no longer be driven back into the bottle. If you want, you can find a company (if necessary abroad) that can carry out any tests.

In another test, the reporter had claimed that he had an affair with a married woman a few years ago and now wants to know whether her daughter is his. Companies are also ready for such problematic tests. There is still a lot of dynamic ahead of us. :-(


April 2011:
The standard reports on DNA tests in immigration procedures in family reunification. Researchers from Finland, Germany and Austria analyze the practice in immigration policy. You find that in all of these countries DNA tests are used today when applying for family reunification, sometimes with unsightly results. A man wants to bring his children (the mother had died in the meantime) to catch up, but the DNA test says that only one of the four children is his, 3 are left behind. Practicing the test broke his "right not to know" about his wife's loyalty. In the laws in D. and Ö. If the respective laws remove the confidentiality of the test results for migrants, the data will be passed on to the police. In Austria "in paragraph 18 of the Aliens Law Amendment Act of 2009, the tests are offered as an option, an option that, according to Interior Minister Maria Fekter, anyone who has nothing to hide will accept." Anyone who does not have a test done for fear of possible findings, worsens their chances of reunification. In contrast to Germany, in Ö. Adopted or fostered children as well as homosexual partners are counted as part of the family. This raises the question of the applicability of the law. In Finland, too, the test is optional, but not decisive.




Back to top

3. Genome analysis by analyzing the real genetic material, e.g. for medical purposes or ancestry tests (genealogy)

How much science is really in these tests?

A slightly shocking test result in 2019: Twins get some 'mystifying' results when they put 5 DNA ancestry kits to the test. Twins submit their identical DNA both to 5 DNA testing companies.

Not only do you get different results from the 5 companies with regard to their genetic origins, the results of the two twins do not match either. Unsettled, they had their DNA tested again by a university - yes, they have identical DNA. The only possible solution is that these tests must have a lot to do with reading coffee grounds.

In the genome analysis (in contrast to the first two techniques) the real, ie the "coding" genetic material (ie the genes that determine the composition of the proteins that are assembled in the cell by the ribosomes) is examined for possible clues to find susceptibility to disease and information about the origin of ancestors (careful interpretation: every person has 4 grandparents, 8 great-grandparents, etc. A mutation at one point in the genetic make-up that is typical for a certain origin only means that there was at least 1 ancestor in the family tree at some point since the creation of mankind who had this mutation.)

The graph shows how much the private DNA testing market is booming in the United States. Much of this data is available for more or less creepy uses.

For such genome analysis, a fully automated test for so-called single nucleotide polymorphism (SNP) is carried out with the aid of DNA chips. These are locations in the genetic material in which a single base pair deviates from the "normal" sequence of base pairs. They are referred to as "successful" point mutations, that is, genetic changes that have prevailed to some degree (1%) in a population's gene pool.

Some SNPs correlate (more or less well) with susceptibility to certain diseases or with certain (regional) ethnic groups. Thanks to the complete automation, such tests can be carried out quite inexpensively (less than 1000 euros). As a result, the customer receives a data record with the values ​​of approx. 100,000 SNPs and an interpretation. The data set is mostly of high accuracy, the interpretation is often problematic, because our knowledge of the effects of certain mutations is very, very rudimentary. What can I do with a statement that my risk of Alzheimer's disease in old age is double, triple or ten times higher than the average?

An article in the NYT in 2012 examined the extent to which diseases can be predicted in this way: Study Says DNA’s Power to Predict Illness Is Limited. Quote:

    If every aspect of a person’s DNA is known, would it be possible to predict the diseases in that person’s future? And could that knowledge be used to forestall the otherwise inevitable? The answer, according to a new study of twins, is, for the most part, "no."

Why are some of these tests very cheap?

Some of these tests are very affordable. This is also easy to explain: As with many services on the Internet, the user is not the customer, but the product. The low prices allow many people to get tested and these test results are resold very profitably: DNA-testing company 23andMe has signed a $ 300 million deal with a drug giant.

The test rate for Americans of European origin is now over 60%. Is it legal to resell? Yes, if the user has agreed to this with the user agreement. However, there is often the option of revoking this. The article linked above explains how to do this.

In this study, 50,000 identical twins were screened for 24 possible diseases. The results show that almost all of them find that their risks are average, e.g. for breast cancer or lung cancer. According to this article, better predictions can be made for the following diseases: Alzheimer's disease, autoimmune thyroiditis, type 1 diabetes and, in men, heart problems. The point is, other than the potential heart problems, there is no way to take any preventive action. And ultimately, the chances are still so small that it is more likely that these people will die from another cause anyway. The results have statistical value and no meaningfulness for the individual case, i.e. they are unsettling, but do not help.

In view of this aspect of limited informative value, Estonia's initiative to carry out a free test for possibly problematic gene variants for around 10% of its citizens is by no means without problems. Citizens can choose which gene variants they want to be informed about. For example, they could decide not to be informed about increased disease risks if there is no preventive measure.

Nevertheless, such an offer naturally puts you under pressure, possibly also from family members who have not received this offer, but are nervous about the inherited risks in their families. Another article in New Scientist poses the problem that too much testing can lead to unnecessary uncertainty.

In addition to the problem that each of these statements only certifies the likelihood of an illness and not that the person will definitely get the illness, there is also the problem that the results of such tests can simply be wrong with a certain probability: The Online Gene Test Finds a dangerous mutation. It may well be wrong. In the case of this article, the person sent the raw results that 23andMe provided to another company who wanted to do a deeper analysis. A susceptibility to illness was diagnosed there. The person knew that there are so-called false positives in every test (i.e. something is supposedly found that is not included in the test material). The person spent a stressful 14 days waiting for the results of a second test by another company. (More on the subject of false positives in my text on face recognition. Fall positives always become a problem if one expression, e.g. terrorist or a rare disease, is very rare in the general population. Then even small error rates of less than 1 % too many misses.

This result is not atypical. The testing companies that carry out the actual test, such as 23andMe, or AncestryDNA, are checked and approved. Other companies such as Promethease in this case only make an additional evaluation and are not subject to any quality controls. A test showed that 40% of the results were incorrect. Most companies protect themselves by not being able to guarantee that the results are correct in the small print. This raises the fundamental question of the quality and informative value of such tests and what questionable results do to the psyche of those affected.




Back to top

Literary treatment of the topic
'Surprise at the DNA test'

Wired brings a book review to Dani Shapiro: Inheritance. The point is that the author's wife has a DNA test carried out, encounters inconsistencies and then researches a long-kept family secret that leads to a proper emotional dynamic.

Such challenges arguably arise with a good number of DNA testers. The question is, do you really want to know all of this?


Finding your father, other relatives, or a murderer on the Internet: genealogical DNA analysis

Google meets DNA database - children can find their biological fathers on the Internet

Report in New Scientist 2005: A child finds his birth father, who was an anonymous sperm donor, via an Internet DNA testing service, although the father has never made his DNA available to this or any other Internet service. The article lists the links of companies that search for people on the Internet on the basis of DNA.

I think we cannot at all yet assess what transparency is in store for us. The donors were promised anonymity. Depending on the respective laws in the respective countries, it could now even be possible that the sperm donors could be called upon to make maintenance payments, which could result in inheritance claims. The article also says at the end that the police could of course search for suspects in this way on the Internet without them ever having given their DNA for testing.

December 2010:
23andMe (and the many other providers) offer "family reunification" for their customers: Find new family with just one click. The extent to which privacy is protected depends on the details of the implementation and how generously the submitters of the DNA share their data. Many of these websites see themselves as a special form of "social networks". Anyone who has submitted their DNA can contact people who have a similar family tree and allegedly close contacts are formed, including visits to the newly found relatives. These can be people with whom contact has been lost, but of course children often also find their biological father (even if they had not looked for it at all, because often they had no doubt about their origin). This can create a powerful dynamic in families and relationships.

Here is the degree of transparency created by such private genealogical ("non-police") DNA databases: In 2018, the number of stored persons is estimated at 12 million US citizens. This will find over 60% of all US citizens with European descent, because it is enough if a third cousin is found there. This is a significant privacy issue.

Feb. 2012:
The NY Times "Ethicist" discussed in the Sunday supplement in 2012 the ethical aspects of medical or parentage tests and whether it is OK if one of several siblings decides for themselves. Because if only 1 person from a family can be tested, the results also apply to some of the other family members, e.g. the siblings, with some cutbacks. This means that if the brother has himself tested and his data becomes known, this also results in possible privacy problems for his siblings, e.g. if a predisposition to an illness is discovered and this information gets to the health insurance company.


Genome analysis and self-chosen ethnic allocation

A very interesting article describes a study by the company 23andme (which is mentioned elsewhere in this part). They also ask their customers who want their genome analyzed, which race they feel they belong to. And then 23andme measures what proportion of typical genes of African, Indian and European origin they find in the respective genome.

It is very important that all these assignments are "self-identified", i.e. the following numbers do not tell anything about the genealogy of so-called African-Americans, but describe how the customers of 23andme classify themselves. Here are some quotes from the article:

"Most Americans with less than 28 percent African-American ancestry say they are white, the researchers found. Above that threshold, people tended to describe themselves as African-American.

On average, the scientists found, people who identified as African-American had genes that were only 73.2 percent African. European genes accounted for 24 percent of their DNA, while .8 percent came from Native Americans.

[Self-identified] Latinos, on the other hand, had genes that were on average 65.1 percent European, 18 percent Native American, and 6.2 percent African. The researchers found that European-Americans had genomes that were on average 98.6 percent European, .19 percent African, and .18 Native American.

Based on their sample, the resarchers estimated that over six million [self-identified] European-Americans have some African ancestry. As many as five million have genomes that are at least 1 percent Native American in origin. One in five African-Americans, too, has Native American roots. "

A contribution by the US broadcaster Radiolab sums up the topic that race is self-assigned very well in this podcast: Allys Choice.

Surprising results and how they are dealt with

In the US, DNA for ancestry clarification (are my ancestors from England, or Italy, or Africa?) Is now so common that there are a significant number of people for whom the result is a significant shock. Relatives give each other these tests, others are just curious, as of 2018, 12 million people had their parentage examined. Companies like 23andMe or AncestryDNA celebrate happy meetings between newly found distant relatives on their websites.

There is even a tendency that social networks can form on the websites of the DNA testers, where people can enthusiastically appreciate their newly found identities and ancestry and discuss them with others of the same ancestry. Even "DNA-based music playlists" will be offered in 2018.

But there are also harrowing realizations: a significant number of people learn that they are not related to their father and this creates considerable stress for them and their families.

The estimates of how many people are affected vary greatly, Wikipedia gives a few percent for our culture. With 12 million tests, the number then adds up to such an extent that self-help groups are now being created on the Internet, and above that is the article linked above. The support group in the article has over 1000 members and there are several such groups.

Reasons for the "surprising" fatherhood can be varied, very often taboos are touched, e.g. simple "cheating", but possibly also rape or incest. A genealogist is quoted in the article with “It's getting harder and harder to keep secrets in our society. If people haven't come to that realization, they probably should. " Nobody expected in the second half of the 20th century that such facts could ever be uncovered with a $ 99 test. The article reports of considerable family stress, often the attempts of the (now mostly adult) children to find the biological father and deep family conflicts and discord.

I dare to doubt whether our society is really so enlightened that it can deal with a situation in which there are no longer any secrets with little conflict. Here technology has opened another Pandora's box.





Here is the family tree of the perpetrator (Suspect) - Source: NY-Times: Technique Used to Find Golden State Killer Leads to a Suspect in 1987 Murders
The two cousings of the perpetrator (identified later) were found in the genealogical database. By searching for common ancestors and excluding people who, for example, live in distant places, the perpetrator was found and then convicted using conventional DNA identification.

Solving crimes using genealogy

May 2018: Great excitement in the US: How a Genealogy Site Led to the Front Door of the Golden State Killer Suspect. With a few controversial steps, a serial killer was found with the help of a genealogy database GEDmatch who had not been active for a long time.

There was a DNA sample from one of the crime scenes, which the police have now submitted as "own" DNA to (allegedly) search for relatives. Although the perpetrator was not found, relatives of his were found who then led to the perpetrator. Here are more details.

A little later, the second case is revealed with the same trick: Technique Used to Find Golden State Killer Leads to a Suspect in 1987 Murders. A graphic in this article shows very nicely that quite distant ancestors of the perpetrator were found and how the police then slowly moved closer. The article says that a number of other cases are still in progress.

These cases show a fundamental problem with DNA publications: The person who sends their DNA to such databases agrees to the processing of their DNA and often / mostly also to the publication on the Internet. However, it also indirectly releases the DNA of its relatives for the search without their consent. Susceptibility to disease also affects not just one person, but often their relatives as well (see next section).


2019: DNA analysis company GEDmatch bought from forensic company

One result of the forensic possibilities of the DNA databases: In 2019, the DNA analysis company GEDmatch will be bought by a company that specializes in supporting police stations.

Most citizens are still involved in solving crimes, but in China DNA analysis is also used to prosecute and monitor minorities and critics.

This is supplemented by DNA phenotyping, i.e. the creation of phantom images from DNA. With the help of the DNA and faces of the comprehensively measured Uighurs, western and Chinese research together on phenotyping and thus enable additional monitoring options. See below for more details on DNA phantom images.


Linking the police DNA databases with the genealogical / medical databases

Since the police authorities began collecting DNA data, the big argument has always been that this data can only be used for 1-to-1 comparisons, i.e. that (apart from the sex of the perpetrator) no statements are made about health, hair color or ethical background can. This has been different since 2018. Therefore, the police in the cases in the USA had to create a false profile, then pass the perpetrator DNA as their own and thus gain access to the offender's relatives.

A study published in 2017 Linkage disequilibrium matches forensic genetic records to disjoint genomic marker sets says: "We show that the method can link a dataset similar to those used in genomic studies with another dataset containing markers used for forensics". The statement of the study, which was hardly noticed until the end of 2018, is that there are links between the non-coding sequences in the police databases and properties such as eye or hair color and health, even without the links in the police databases that now appear possible is implemented with the genealogical databases.

The experts and the users of the genealogical databases are arguing about whether they have a problem with collecting criminals with the help of DNA data given for other purposes.


Secret tests for susceptibility to disease

From a privacy perspective, it becomes very problematic if these tests are not carried out for one's own genome but in secret for other people without their consent. As mentioned above, this is forbidden in Great Britain, but the testers from New Scientist managed to have a colleague's genetic material tested for possible susceptibility to disease. The very interesting article from 2009 shows how difficult it is to really keep private things private in the "brave new world" of biotechnology.

Who could be interested in testing another person's genetic makeup? There are, for example, employers who want to know how high the probability of various types of cancer or the like is. Of course, life insurers are also very interested in such information. In both cases it does not bother at all that the statements obtained in this way are only probability statements, but with a large number of applicants, employees or potential insurance customers, there are still valuable opportunities for optimization. But also in the private sphere, people with marriage plans could be interested in the health risks of their potential spouse. Or potential parents could secretly get a DNA test of the child prior to adoption.

You can find out how the DNA of strangers can be secretly tested

    The challenge for the reporter from New Scientist was that these tests using a DNA chip require much more DNA than for an identity or ancestry determination. Very little DNA can be extracted from a used drinking glass. So he first had to find a company that would carry out a so-called "whole genome amplification" (without asking him to provide evidence that this was his own DNA). Such a company has been found and duplicated the DNA sample from the drinking glass for him.
    With the sample obtained in this way, he turned to 3 companies abroad: Decode Genetics of Reykjavik, Iceland, and 2 companies in California 23andMe and Navigenics. The two companies in the USA ask for 2 milliliters of saliva for the test, but the reporter could not get it.
    The company in Iceland, on the other hand, sends a piece of filter paper on a handle to be rubbed on the inside of the cheek. The reporter took the colleague's genetic material that he had copied and applied it to the filter paper. However, this sample gave no result with the DNA chip used in Iceland. In a second attempt, the reporter got a used condom from his colleague and prepared the filter paper with the semen. This then worked - this method seems to be quite suitable if a woman wants to have the genetic material of her future child tested.
    In a final experiment, the reporter had the duplicated DNA from the water glass tested with another DNA chip and received the desired data, but in this case without interpreting the data. But this is not a problem. The free program Promethease can be found on the Internet on the Genomics website This program is supervised by enthusiasts who constantly update the research results with correlations between diseases (or regional ethnic groups) and the SNPs. With its help it was then easy to interpret the genome data.
    The precision of the raw data without prior duplication of the DNA is very high (99.996%), with the duplicated DNA only 93%. The interpretation of the data is (currently) still largely a coffee grounds forecast.
    The reporter showed that even in Great Britain, where testing of someone else's DNA is banned, it is not a great challenge to get someone else's DNA and have it analyzed. Reproduction would not even have been necessary for a paternity test.




Back to top

What happens to the raw data, the submitted DNA?

The medical tests only look at certain genes, but of course all of the DNA is submitted. The US authorities have now begun to make claims on this raw data: They are looking for DNA that is somehow "similar" to the one at the crime scene. If they find something like that, then someone from the family of the test person could have been the perpetrator and then the family has the problem that they somehow have to come out of suspicion again. How subjective such "matches" can be is explained above.

November 2011:
After intense pressure on the Austrian authorities, the US authorities are given direct access to the Austrian DNA database. The US is said to have access to DNA databases, fingerprint files and the identities of terrorist suspects based on an agreement that was signed under intense American pressure. The agreement gives the US authorities access to the domestic police databases. Specifically, US authorities should be able to make inquiries to Austria as to whether a secured DNA profile or a fingerprint is registered in Austria. However, the Austrian files cannot differentiate between light and serious crimes, i.e. shoplifters can also be found this way.

Jan. 2015:
23andMe has landed a big financial coup: They sell their customers' DNA data (with their blanket approval in the terms of use) to Genentech for up to $ 60 million, which they want to use for their research. Another example where the users (who only pay $ 99 themselves) are not the company's customers, but the product that the company is marketing. This is the new "data economy".




Phantom images from DNA - Forensic DNA Phenotyping

In the process of establishing identity for forenic purposes, which has been described so far, only non-coding sections from the so-called junk DNA are used. These sections can tell with a certain probability whether 2 DNA problems originate from the same human (or animal). They do not allow any further statements about the person, i.e. of all DNA analyzes they are the least drastic from the point of view of privacy. But the desires of forensic scientists (and other interested parties) don't stop there.

A new development in 2014: Scientists claim that they can not only provide clues about a possible genetic origin from the DNA, they can even estimate what the person's face looks like with a good probability: To Sketch a Thief: Genes Draw Likeness of Suspects.

Scientifically this is called "DNA forensic phenotyping". It started in 2004 with a few cases where DNA was used to determine ethnic origin. However, since this relatively simple task is already fraught with many errors ("false positives"), the idea of ​​generating complete phantom images is met with considerable skepticism.

March 2015:
A year later it looks like progress: Building a Face, and a Case, on DNA.

The comparison between photo and DNA-generated image in 2015 is similar (or dissimilar)

The article describes that ultimately an attempt is made to estimate the percentage of African, European and Asian DNA contained in the DNA. And from this an average picture of this mixture is generated. Age and gender are also taken into account. Hair color could be supplemented, but doesn't say much after coloring. That is, the resemblance to the average person with these ancestors is very limited.

For legal reasons, in some US states other findings that could be obtained from the DNA, such as whether there is a predisposition to certain diseases or the like, are (currently) not taken into account. On the other hand, the DNA that is found at the scene of a crime is "abandoned material" and does not enjoy legal protection in the USA.

2015: Germany and Belgium currently do not allow forensic DNA phenotyping and in the Netherlands it is restricted to externally visible attributes such as hair and eye color, not to diseases or other findings from the DNA.

2018:Germany wants to legalize DNA phenotyping for police purposes. And that without storage periods and without any restrictions on the purpose of DNA storage.

2019: There are advances in DNA phenotyping. With the help of the DNA and faces of the comprehensively measured Uighurs, western and Chinese research together on phenotyping and thus enable additional monitoring options.




Back to top

Further information

May 2010:
The ORF reports on a study by the Austrian Bioethics Commission on gene and genome tests on the Internet (pdf).

Jan. 2011:
The NY Times reports a study by The New England Journal of Medicine on the effects of DNA testing on health risks on testers. The study is also directly on the Internet: Effect of Direct-to-Consumer Genomewide Profiling to Assess Disease Risk.


Philipp Schaumann,


Copyright notice:
If no other author is named, the copyright of the material on this website lies with Philipp Schaumann.
These texts are licensed under the Create Commons Attribution-Noncommercial-Share Alike 2.0 Austria license. Of course, the rules of fair use also apply and exceptions to the license can be discussed with the authors at any time.