
Genetic testing toy
“Welcome to yourself”, “get to know your DNA”, “explore your inner universe” - these are the slogans that make up companies offering genetic testing. Indeed, it is an amusing thing to look at the failures of your own genome or find out where your genus originates from. Now it is quite simple to do this, because IT companies were able to “wrap” the results of biological research in a form that is understandable to us. I spat in a test tube today, and after a month you already see multicolored reports on the site.

A year ago, I used the services of 23andMe , the largest global company that provides genetic services directly to the client (direct-to-customer). Using the example of my result, I would like to tell you what is hidden behind such a genetic test.
We will deal with the biological part of the question. The human genome is a complete set of genetic information, which, for the most part, is packaged in 23 pairs of chromosomes. The human genome has a whopping 3.2 billion pairs of DNA bases (A, C, G or T). It is as if all the people on the planet have lined up in a strongly curved curve in pairs, and we have two such lines: from mom and dad. In addition, in each of the 3.7 trillion cells in our body, the genome is the same. That is, if
some change has crept into our genetic code, then it’s impossible to “encode” it yet.
However, from a genetic point of view, all people are twins, because the genomes of two people are similar to 99.9%. Moreover, the differences are mainly represented by point mutations, when one nucleotide changes to another. So, several point mutations can distinguish a curly-haired person from a person with straight hair, healthy from a cancer patient.
Unfortunately, we do not know as much as we would like about the connection between genome symbols and visible characters. This is due to the complexity and variety of mechanisms that occur in cells. For a number of signs, this relationship is well shown, for example, for such as blood type, eye color, or susceptibility to breast cancer.
Mutations are bio-markers targeted by 23andMe and several other companies. Now their genetic test takes into account 1 million mutations in the genome. However, these are not all possible mutations. It is worth noting that there are tests that target completely different markers. For example, paternity tests analyze short repeats of specific sections of DNA. These sites vary widely between families, but are transferred almost without change within the same family.
In order to do a genetic test, you need to send a sample of your DNA. 23andMe and many other similar companies isolate DNA from saliva. It is worth noting that saliva is not the best material for DNA analysis, since the cells that it contains, like other somatic cells, can accumulate random mutations. I will not write what material and why is more suitable for DNA analysis.

The process is highly automated. You are sent a small box with a test tube that you need to fill with saliva to the line. The test tube has a special nozzle to make it easy to fill. Upon closing, the film on the nozzle breaks, and the tube is filled with buffer solution so that the sample does not deteriorate upon delivery. Then you seal it and ship it to the USA with DHL. You also need to enter a unique tube number on the 23andMe website, which then matches your DNA sample and your account. In semi-automatic mode, the bio-laboratory detects mutations and sends 23andMe data. That is, the laboratory itself does not draw any conclusions, but only provides raw data on mutations. The entire analysis is carried out with bioinformation tools. The laboratory also does not know the names, but sees only the unique numbers of the tubes.
23andMe test can be ordered in almost all countries of the world, but not in Russia. Because in our country, unauthorized transportation of biomaterial is illegal. There are attempts to find loopholes in this system, but they either failed, or ate too much time and nerves. I received and sent my test from Italy.
The cost of the 23andMe test is $ 99 plus shipping. For residents and visitors of the United States, delivery will cost about $ 5, for Europeans it is about $ 60.
In large cities of Russia there are similar services and even 23andMe clones, but their prices are much higher, and the quality and volume of analyzes are lower so far.
Before doing genetic tests, you need to decide on ethical issues. Firstly, genetic diseases are quite insidious and many of them are still incurable. You may be a carrier of a mutation that can lead to phenylketonuria. You can not get sick with phenylketonuria, because a copy of the same gene from another parent is normal, but your offspring may be susceptible to this disease. Or a test may show you that you have a very high risk of developing Alzheimer's. This disease is characterized by impaired mental activity and usually manifests itself after 50 years. Living with or without such information is a purely individual matter.
Secondly, to some extent, genetic information is common to the whole family. That is, if you are a carrier of some kind of mutation, then most likely it came from one of your parents and it will also be transmitted to your descendants. And the test result will affect not only you, but also your entire family.
Thirdly, individual genetic information is a new phenomenon, and it is not yet clear how to relate to its confidentiality. 23andMe protects your data pretty well, but you can open it yourself. Undoubtedly, the more data we have in the public domain, the more power our statistical conclusions have. But who knows how this will be used in the future. For example, in America, genetic data can now affect customer insurance rates.
The 23andMe website gives you visual reports of test results. Everything is convenient, systematic and simple.
I identified 4 main categories of information received. The list is sorted in descending order of importance to me:
Genealogy tests are based on large-scale studies of full or partial genomes taken from different populations. Scientists were able to identify sets of mutations that are characteristic only for a given population (North Americans, Chinese, Australians, etc.). Having a set of mutations, you can determine to which groups and in what percentage the person belongs. Such tests have pretty good accuracy.

The graph shows my belonging to the East European population. Since 23andMe does not work with Russia, the population closest to it is displayed. In principle, the result is true, because my parents are from Russia. It was interesting to find a small percentage of Yakut and Balkan roots. There were no Italians or French at all in my family tree, which killed my attempts to explain my love of wine.
Well-studied genetic basis of a not too wide range of diseases. For example, for diseases such as phenylketonuria, breast cancer, Bloom's syndrome, hemochromatosis and others, mutations are known that can cause them. The test shows whether you are a carrier of one of these mutations or not. Their presence can lead to serious risks of developing diseases. But mutations can also be “silent,” that is, manifest only in future generations. 23andMe gives an answer “yes or no” for about fifty diseases. I found only one “silent” mutation, which can lead to hemochromatosis - the accumulation of iron in the body. Judging by the 23andMe test, there are no abnormalities in my genome. This may mean that my genome is clean. And it may mean that the remaining dangerous mutations have not yet been discovered by science.
If the genealogical and hereditary information looks very convincing, then I am skeptical about the risks and other tools of 23andMe.
23andMe has its own disease risk assessment system. This system covers both open research and databases, as well as their own results. Risk ratings are sorted in descending order of 23andMe confidence in this result. Four stars of confidence mean that the result is based on at least two studies in which at least 750 subjects were involved. Three or less confidence stars mean that the result is based on initial research on illness and mutation. The report on my genome is not an abundance of disease.

Risks are divided into three groups: increased, decreased, and normal (average for the population) risk. If the risk assessment for the disease is higher than the average, the disease falls into the high-risk group. You can see a description of each disease, what it threatens with, and on the basis of what studies this risk has been assessed. It also provides recommendations on what to do to reduce the risks of the disease. Such tips are general in nature: do sports, do not eat a lot of salt, consult a doctor and so on.
In addition to assessing health risks, there are other 23andMe cognitive tools. Star-based confidence ratings are based on the digestibility of certain medications. In my report, only drugs that I had never heard of differed 4 stars of confidence: warfarin and abacavir. In another report, 23andMe tried to predict my external characteristics (eye color, prone to baldness and others) from my genomic data. The information in this report was about half true. It was interesting to know that my genome is 3% similar to the Neanderthal genome. Our ancestors crossed with Neanderthals, which left a mark on our genome. A relative search tool found some six-cousins from the United States.
I did not find the formulas by which 23andMe risk assessments are calculated. Apparently, they have not been published, as well as the database of test results of 650 thousand of their customers. So I rummaged through the technical information to check how accurate the risk assessments are.
The highest risk disease in my report is venous thrombosis. Due to the increased density, blood can form a blood clot in the veins, which causes pain and interferes with the free flow of blood. As a rule, blood accumulates in the veins of the legs. 23andMe reports that the average risk of developing this disease is about 12%, and my risk is approximately 36%.
The assessment of this risk is based on three mutations (identifiers rs6025, rs1799963 and rs505922). This information can be found in the technical report, for each mutation a number of articles are presented with a study of the relationship between this mutation and the disease.
23andMe provides the ability to download raw data. The file contains all the mutations that have been tested at the client. I downloaded my file and opened it using SNP Toolbox. With this program you can see all the mutations and genes into which they fall, as well as see the relationship of mutations and diseases based on information from open sources. We remember that we get two different genomes - from mom and dad. The SNP Toolbox program showed that the first mutation has the CC genotype, that is, in this position of the genome in one version of this gene (for example, from mom) costs C, and in another version (from dad) also costs C. Genotypes AG, AA, GG positively or negatively associated with venous thrombosis. CC and GG in this case are the same genotype, since C and G are complementary nucleotides. According to open sources, the GG genotype (CC) does not cause a risk of venous thrombosis. The second mutation was indeed associated with venous thrombosis, but there is no numerical risk indicator in open sources. The third mutation is in the non-coding part of the gene, and there are no known clinical studies on it. As it turned out, information about the third mutation came from genome-wide studies on the association of mutations and diseases (Genome-wide association studies), which are based on statistical data, and not on clinical trials.
Next, I looked at the publications from the 23andMe report. Basically, these were publications of the late 90s - early 2000s. This is research in the so-called “pregenomic” era - before the decoding of the human genome in the 2000s. Among these publications, there was only one large-scale study with 5.5 thousand subjects. This study is mentioned in the report on the first and second mutations, apparently, the system considers it two different. The aim of this study is to assess the risk of venous thrombosis in carriers of both of these mutations. However, only one of them is present in my genotype, that is, the results of this study are not applicable to me.
I did not dig further, since the result was already enough to doubt the accuracy of the 23andMe estimates. Moreover, I still do not understand where the numerical values of these estimates come from.
Those who want to do 23andMe analysis now should know that the Food and Drug Administration (FDA) has banned the company. Litigation is ongoing. The FDA believes that a program that advises clients based on their genetic information is a medical device. And such devices must be certified. Due to the FDA ban, only genealogy information and raw data are available to 23andMe new customers. Any health information is not yet available.
You can work with 23andMe raw data outside the company's website. For example, you can upload them to the openSNP database, a project for collecting open access genomic data. It provides virtually no useful information to the downloadable, but helps scientists who study genomic variations. The site has already collected more than 1000 test results.
Of the 21 sites that work with 23andMe data, I managed to launch only 2 on the go and for free.
The first of them showed that 4 mutations in my genome are somehow related to methylation. This information is not of particular importance to me.
The second site I managed to work with is Genes and Usturned out to be much more interesting. From the 23andMe tests of two different people, the site derives estimates of the health risks of their possible offspring. Since I did not have the results of the second test, I tried to look at potential offspring with unsuspecting Susan Smith, a demo. It turned out that our potential offspring has an increased risk of contracting venous thrombosis and prostate cancer. However, this result is based on two inaccurate estimates, so I don’t think it’s worth it to write off Susan Smith.

A year ago, I used the services of 23andMe , the largest global company that provides genetic services directly to the client (direct-to-customer). Using the example of my result, I would like to tell you what is hidden behind such a genetic test.
Mutations
We will deal with the biological part of the question. The human genome is a complete set of genetic information, which, for the most part, is packaged in 23 pairs of chromosomes. The human genome has a whopping 3.2 billion pairs of DNA bases (A, C, G or T). It is as if all the people on the planet have lined up in a strongly curved curve in pairs, and we have two such lines: from mom and dad. In addition, in each of the 3.7 trillion cells in our body, the genome is the same. That is, if
some change has crept into our genetic code, then it’s impossible to “encode” it yet.
However, from a genetic point of view, all people are twins, because the genomes of two people are similar to 99.9%. Moreover, the differences are mainly represented by point mutations, when one nucleotide changes to another. So, several point mutations can distinguish a curly-haired person from a person with straight hair, healthy from a cancer patient.
Unfortunately, we do not know as much as we would like about the connection between genome symbols and visible characters. This is due to the complexity and variety of mechanisms that occur in cells. For a number of signs, this relationship is well shown, for example, for such as blood type, eye color, or susceptibility to breast cancer.
Mutations are bio-markers targeted by 23andMe and several other companies. Now their genetic test takes into account 1 million mutations in the genome. However, these are not all possible mutations. It is worth noting that there are tests that target completely different markers. For example, paternity tests analyze short repeats of specific sections of DNA. These sites vary widely between families, but are transferred almost without change within the same family.
Process
In order to do a genetic test, you need to send a sample of your DNA. 23andMe and many other similar companies isolate DNA from saliva. It is worth noting that saliva is not the best material for DNA analysis, since the cells that it contains, like other somatic cells, can accumulate random mutations. I will not write what material and why is more suitable for DNA analysis.

The process is highly automated. You are sent a small box with a test tube that you need to fill with saliva to the line. The test tube has a special nozzle to make it easy to fill. Upon closing, the film on the nozzle breaks, and the tube is filled with buffer solution so that the sample does not deteriorate upon delivery. Then you seal it and ship it to the USA with DHL. You also need to enter a unique tube number on the 23andMe website, which then matches your DNA sample and your account. In semi-automatic mode, the bio-laboratory detects mutations and sends 23andMe data. That is, the laboratory itself does not draw any conclusions, but only provides raw data on mutations. The entire analysis is carried out with bioinformation tools. The laboratory also does not know the names, but sees only the unique numbers of the tubes.
23andMe test can be ordered in almost all countries of the world, but not in Russia. Because in our country, unauthorized transportation of biomaterial is illegal. There are attempts to find loopholes in this system, but they either failed, or ate too much time and nerves. I received and sent my test from Italy.
The cost of the 23andMe test is $ 99 plus shipping. For residents and visitors of the United States, delivery will cost about $ 5, for Europeans it is about $ 60.
In large cities of Russia there are similar services and even 23andMe clones, but their prices are much higher, and the quality and volume of analyzes are lower so far.
Ethics
Before doing genetic tests, you need to decide on ethical issues. Firstly, genetic diseases are quite insidious and many of them are still incurable. You may be a carrier of a mutation that can lead to phenylketonuria. You can not get sick with phenylketonuria, because a copy of the same gene from another parent is normal, but your offspring may be susceptible to this disease. Or a test may show you that you have a very high risk of developing Alzheimer's. This disease is characterized by impaired mental activity and usually manifests itself after 50 years. Living with or without such information is a purely individual matter.
Secondly, to some extent, genetic information is common to the whole family. That is, if you are a carrier of some kind of mutation, then most likely it came from one of your parents and it will also be transmitted to your descendants. And the test result will affect not only you, but also your entire family.
Thirdly, individual genetic information is a new phenomenon, and it is not yet clear how to relate to its confidentiality. 23andMe protects your data pretty well, but you can open it yourself. Undoubtedly, the more data we have in the public domain, the more power our statistical conclusions have. But who knows how this will be used in the future. For example, in America, genetic data can now affect customer insurance rates.
Test result
The 23andMe website gives you visual reports of test results. Everything is convenient, systematic and simple.
I identified 4 main categories of information received. The list is sorted in descending order of importance to me:
- Genealogical origin;
- Diseases and characteristics with a studied genetic basis: there is a mutation / no mutation;
- Risks assessed by 23andMe;
- Cognitive.
Genealogy tests are based on large-scale studies of full or partial genomes taken from different populations. Scientists were able to identify sets of mutations that are characteristic only for a given population (North Americans, Chinese, Australians, etc.). Having a set of mutations, you can determine to which groups and in what percentage the person belongs. Such tests have pretty good accuracy.

The graph shows my belonging to the East European population. Since 23andMe does not work with Russia, the population closest to it is displayed. In principle, the result is true, because my parents are from Russia. It was interesting to find a small percentage of Yakut and Balkan roots. There were no Italians or French at all in my family tree, which killed my attempts to explain my love of wine.
Well-studied genetic basis of a not too wide range of diseases. For example, for diseases such as phenylketonuria, breast cancer, Bloom's syndrome, hemochromatosis and others, mutations are known that can cause them. The test shows whether you are a carrier of one of these mutations or not. Their presence can lead to serious risks of developing diseases. But mutations can also be “silent,” that is, manifest only in future generations. 23andMe gives an answer “yes or no” for about fifty diseases. I found only one “silent” mutation, which can lead to hemochromatosis - the accumulation of iron in the body. Judging by the 23andMe test, there are no abnormalities in my genome. This may mean that my genome is clean. And it may mean that the remaining dangerous mutations have not yet been discovered by science.
If the genealogical and hereditary information looks very convincing, then I am skeptical about the risks and other tools of 23andMe.
23andMe has its own disease risk assessment system. This system covers both open research and databases, as well as their own results. Risk ratings are sorted in descending order of 23andMe confidence in this result. Four stars of confidence mean that the result is based on at least two studies in which at least 750 subjects were involved. Three or less confidence stars mean that the result is based on initial research on illness and mutation. The report on my genome is not an abundance of disease.

Risks are divided into three groups: increased, decreased, and normal (average for the population) risk. If the risk assessment for the disease is higher than the average, the disease falls into the high-risk group. You can see a description of each disease, what it threatens with, and on the basis of what studies this risk has been assessed. It also provides recommendations on what to do to reduce the risks of the disease. Such tips are general in nature: do sports, do not eat a lot of salt, consult a doctor and so on.
In addition to assessing health risks, there are other 23andMe cognitive tools. Star-based confidence ratings are based on the digestibility of certain medications. In my report, only drugs that I had never heard of differed 4 stars of confidence: warfarin and abacavir. In another report, 23andMe tried to predict my external characteristics (eye color, prone to baldness and others) from my genomic data. The information in this report was about half true. It was interesting to know that my genome is 3% similar to the Neanderthal genome. Our ancestors crossed with Neanderthals, which left a mark on our genome. A relative search tool found some six-cousins from the United States.
Own analysis
I did not find the formulas by which 23andMe risk assessments are calculated. Apparently, they have not been published, as well as the database of test results of 650 thousand of their customers. So I rummaged through the technical information to check how accurate the risk assessments are.
The highest risk disease in my report is venous thrombosis. Due to the increased density, blood can form a blood clot in the veins, which causes pain and interferes with the free flow of blood. As a rule, blood accumulates in the veins of the legs. 23andMe reports that the average risk of developing this disease is about 12%, and my risk is approximately 36%.
The assessment of this risk is based on three mutations (identifiers rs6025, rs1799963 and rs505922). This information can be found in the technical report, for each mutation a number of articles are presented with a study of the relationship between this mutation and the disease.
23andMe provides the ability to download raw data. The file contains all the mutations that have been tested at the client. I downloaded my file and opened it using SNP Toolbox. With this program you can see all the mutations and genes into which they fall, as well as see the relationship of mutations and diseases based on information from open sources. We remember that we get two different genomes - from mom and dad. The SNP Toolbox program showed that the first mutation has the CC genotype, that is, in this position of the genome in one version of this gene (for example, from mom) costs C, and in another version (from dad) also costs C. Genotypes AG, AA, GG positively or negatively associated with venous thrombosis. CC and GG in this case are the same genotype, since C and G are complementary nucleotides. According to open sources, the GG genotype (CC) does not cause a risk of venous thrombosis. The second mutation was indeed associated with venous thrombosis, but there is no numerical risk indicator in open sources. The third mutation is in the non-coding part of the gene, and there are no known clinical studies on it. As it turned out, information about the third mutation came from genome-wide studies on the association of mutations and diseases (Genome-wide association studies), which are based on statistical data, and not on clinical trials.
Next, I looked at the publications from the 23andMe report. Basically, these were publications of the late 90s - early 2000s. This is research in the so-called “pregenomic” era - before the decoding of the human genome in the 2000s. Among these publications, there was only one large-scale study with 5.5 thousand subjects. This study is mentioned in the report on the first and second mutations, apparently, the system considers it two different. The aim of this study is to assess the risk of venous thrombosis in carriers of both of these mutations. However, only one of them is present in my genotype, that is, the results of this study are not applicable to me.
I did not dig further, since the result was already enough to doubt the accuracy of the 23andMe estimates. Moreover, I still do not understand where the numerical values of these estimates come from.
FDA
Those who want to do 23andMe analysis now should know that the Food and Drug Administration (FDA) has banned the company. Litigation is ongoing. The FDA believes that a program that advises clients based on their genetic information is a medical device. And such devices must be certified. Due to the FDA ban, only genealogy information and raw data are available to 23andMe new customers. Any health information is not yet available.
Other sites
You can work with 23andMe raw data outside the company's website. For example, you can upload them to the openSNP database, a project for collecting open access genomic data. It provides virtually no useful information to the downloadable, but helps scientists who study genomic variations. The site has already collected more than 1000 test results.
Of the 21 sites that work with 23andMe data, I managed to launch only 2 on the go and for free.
The first of them showed that 4 mutations in my genome are somehow related to methylation. This information is not of particular importance to me.
The second site I managed to work with is Genes and Usturned out to be much more interesting. From the 23andMe tests of two different people, the site derives estimates of the health risks of their possible offspring. Since I did not have the results of the second test, I tried to look at potential offspring with unsuspecting Susan Smith, a demo. It turned out that our potential offspring has an increased risk of contracting venous thrombosis and prostate cancer. However, this result is based on two inaccurate estimates, so I don’t think it’s worth it to write off Susan Smith.