Table of contents:
Video: We Don't Yet Know What Most Of The Genetic Apparatus Is Responsible For
We don't yet know what most of the genetic apparatus is responsible for
The main freelance specialist in medical genetics of the Ministry of Health of Russia, President of the Association of Medical Geneticists of Russia, Corresponding Member of the Russian Academy of Sciences, Director of the Federal State Budgetary Scientific Institution "MGNTs" told Mednovosti about why domestic BigData databases are needed, and how artificial intelligence can improve the situation with the diagnosis and treatment of genetic diseases. d.m.s. Sergey Kutsev.
Photo: pixabay.com /
Rapidly developing digital technologies are increasingly penetrating medicine. The main freelance specialist in medical genetics of the Ministry of Health of Russia, President of the Association of Medical Geneticists of Russia, Corresponding Member of the Russian Academy of Sciences, Director of the Federal State Budgetary Scientific Institution "MGSC" d.m.s. Sergey Kutsev.
Sergei Ivanovich, did artificial intelligence (AI) really revolutionize medical genetics, transferred it from the field of pure science to clinical practice
- Modern genetics operates with a huge amount of data obtained during DNA diagnostics. As a result of testing, we find in the patient's genome changes in the nucleotide sequence, which can be both normal variants (polymorphisms) and the cause of the development of the disease, that is, mutations. And to understand what we are dealing with, it takes analysis of a large number of different databases.
But for now, the use of AI for us is still only scientific work. To use it in practice, it is necessary to standardize large amounts of clinical data. When we have electronic medical records with a standardized description of the clinical manifestations of each patient, we will be able to identify risk groups and specifically examine them. These studies are especially important for finding those diseases for which there is a pathogenetic treatment.
Today, our Medical Genetic Center and other centers have accumulated a large array of information about hereditary diseases and their molecular genetic causes. In some diseases, it is possible to identify the most frequent or major mutations. The diagnostic algorithms created on their basis are included in the clinical guidelines. And of course, it would be good to collect all the data concerning the molecular causes of various hereditary diseases into a single information system. Ideally, every patient should have an electronic card and be available to doctors of all specialties. For example, Israel already has such a system: wherever the patient is being treated at the moment, the doctor has access to his medical history, where the results of all previous examinations are saved.
And, of course, we need a domestic database containing information about certain variants in the genetic apparatus. Analysis of information using Big Data can replace complex additional laboratory research. Today we work with international databases. The most famous of these is OMIM, a database on hereditary diseases, which contains information about the clinical picture and the genetic variants that cause it. As a result of genome-wide sequencing, we get a huge amount of so-called "raw" patient data, which can weigh more than 100 GB. And the software helps to filter out those options that may be causing the disease. But even after automatic processing, there are hundreds of options that are analyzed by specialists.
Why do we need a domestic database when there are already international ones?
- The fact is that the variants of the norm and mutations, as well as the frequency of their occurrence, differ in different populations. Population databases are based on the data of complete exomes and complete genomes of clinically healthy people, and knowing the frequency of a variant in a population, it is already possible to draw a conclusion about what we are dealing with: a norm or a pathology.
For example, now more and more information is received about hereditary forms of epilepsy. In one of our patients, we found a variant that, according to our data, is very rare in the population, and may be both the norm and the cause of epilepsy. We turned to our Chinese colleagues with the question of how often such a change in DNA occurs in them. If it is as rare or occurs in patients with a similar clinical picture, then it is most likely to be the cause of the disease. It turned out that about 5% of the Chinese population have this option, and much fewer people suffer from epilepsy. This means that this option is absolutely not the cause of the disease. It turns out that a simple analysis of information using Big Data can replace complex additional laboratory research.
In addition to our own database, we also need a single platform for all geneticists to work with patients, and for the future analysis of electronic medical records.
Geneticists need the so-called "deep phenotyping", that is, a thorough study of the entire set of signs and clinical symptoms in a patient. Despite the availability of technologies that allow us to study the structure of the genome in great detail, the phenotype in our country is described rather primitively. Nevertheless, even small signs are very often important for diagnosis. For example, according to the characteristics of the face, a good geneticist can immediately suspect this or that syndrome.
What will deep phenotyping give to geneticists and their patients?
- In theory, deep phenotyping includes about 13 thousand features for which it is desirable to examine a patient. All this is Big Data that really needs to be analyzed. If we had such an opportunity and a complete genome, we could do a lot not only for hereditary, but also for multifactorial diseases.
Today, the study of the genome and the study of its diversity is one of the most important directions, and this is a task for the next 20-30 years. Previously, it was believed that the coding part of the genome is 1-2% of the entire genetic apparatus, and all other regions do not play any role. But it turned out that this is not at all the case: these regions encode different classes of RNA, and these RNAs, in turn, are regulators of the function of other genes.
The study of the structure of the genome in health and disease is very actively involved in the whole world. For example, in Great Britain there was a program called "One Hundred Thousand Genomes", which was supposed to study 100 thousand inhabitants of the country. Now the program is called "a million genomes", and in the future will form the basis of health care in the UK. Similar projects are underway in other countries such as China, the United States and others. Unfortunately, we do not have such large-scale projects.
How Big Data can improve the situation with the diagnosis and treatment of rare genetic diseases?
- All hereditary diseases are rare, but some of them are relatively common, for example, phenylketonuria or cystic fibrosis. There are very different approaches to developing therapy. So, genome editing is a direct way of treatment, the correction of a certain defect that we can identify. But Big Data allows you to search for other workarounds as well. In particular: to use gene regulators and search for genes participating in molecular networks, to influence the elements of these networks in order to neutralize the action of another gene or, on the contrary, activate it.
If we have a sufficiently detailed description of the clinical manifestations, approaching what we call "deep phenotyping," we can try to identify risk groups and then more accurately find patients with diseases for which a treatment can be found.
Do you already have specific examples?
Sure. For example, we had an idea to analyze electronic medical records of patients from a high risk group for orphan diseases. Thus, Fabry's disease is multifaceted and the pathological process involves the cardiovascular system, excretory system (kidneys) and the nervous system, which is why patients can get to specialists of different profiles, who can not always come to an assumption about such a rare disease. In our country, there are medical institutions that have electronic patient records. And we analyzed a large number of patient records that have long been observed by nephrologists, neurologists or cardiologists and have clinical manifestations that may be an integral part of Fabry disease. As a result, a group of several thousand patients was formed, which we can call the risk group for this disease.The next step will be screening (there are methods that allow you to determine the activity of the enzyme, the lack of which leads to this disease). And so we will find out how much the artificial intelligence used to analyze the symptoms in electronic maps has helped us.
Can genome-wide sequencing be considered an effective tool for detecting hereditary diseases?
- Without a doubt. This is really what revolutionized our medical genetic service. For example, there are several hundred genes that can lead to epilepsy. And it will be costly and difficult to sequence each gene separately. This will be done later to validate the variants identified in whole genome studies.
Today, it is very popular to create panels when several tens or hundreds of genes are analyzed that are responsible for a particular pathology. It is very important that this technology makes it possible to identify genetic heterogeneity that is characteristic of hereditary diseases. The point is that the same disease can be caused by a mutation in different genes, which means that the treatment can also differ based on knowledge of the genotype. In particular, phenylketonuria can be caused not only by a mutation in the phenylalanine hydroxylase gene: a similar clinical picture can be observed in the presence of mutations in other genes, for example, in hyperphenylalaninemia with tetrahydrobiopterin deficiency.
Finding the only variant that led to the development of the disease is always difficult, because in different regions of each genome there are several hundred thousand variants, and some of them are unique and have not been described previously. Therefore, there is also an approach in which the genomes of the father and mother are also examined and compared with the genome of the child.
This year, a federal program for the development of genomic technologies was adopted. Three centers for genomic research are being established. Will they be engaged in, among other things, population genomics, genetic certification of the population?
- As far as I know, no. One of the three centers will work in the field of biosafety and will deal with infectious pathologies and genomics of microorganisms, the second in the field of agriculture, and the third in the field of medicine. The center will be called Precision Genome Editing and will focus exclusively on the development of genome editing technologies. This is one of the promising areas for the development of technologies for the treatment of hereditary, oncological, autoimmune and infectious diseases (including HIV). But globally, this program does not concern questions of medical genetics. Unfortunately, we have medical genetics aside from federal initiatives.
Editing the genome is not the same as researching the epidemiology of hereditary diseases, the structure of the genome, or the functions of one or another of its loci. And we do not yet know what most of the genetic apparatus is responsible for. Genome editing technology is promising, but not yet applicable for the treatment of hereditary diseases. Thus, side effects are a potential limitation: in addition to the fact that the target (a particular gene) is being edited, off-target editing of other genes also occurs. This can lead, for example, to cancer. But technology is evolving and new tools are reducing the risks of off-target editing.
In Russia, by 2025, they plan to provide all residents with genetic passports. Will this help reduce the risk of having children with hereditary diseases?
The genetic certification that the government is talking about has nothing to do with medical genetics. It is planned to create a passport with certain markers that make it possible to identify a person, and this is a problem of forensic medicine.
There is another understanding of the genetic passport, already related to medicine. Such a passport is not fully scientifically substantiated and suggests, using genetic technologies, to study the predisposition of a person to frequent socially significant diseases, such as diabetes mellitus, bronchial asthma, arterial hypertension, etc. But these are not hereditary, but multifactorial diseases in which a genetic predisposition exists, but the disease develops only under the influence of external factors. Genetic signs of predisposition cannot predict the development of the disease and, in essence, the odds remain 1 to 1.
Another understanding of the genetic passport involves testing each individual for the carrier of pathogenic variants of frequent hereditary diseases. This will allow, when planning childbirth, to compare the passports of the spouses and find out if there is a risk of having a child with one or another mutation.
It seems that this risk is increasing today. Have there really been more gene mutations lately, or have they just been noticed more often?
- The frequency of hereditary diseases remains at the same level, these are the laws of population dynamics. We receive more and more information about such diseases, but this is due to the fact that some of them pass from the non-hereditary group to the hereditary group. Mutations can be different and can be different in different populations, but the percentage of people with hereditary diseases remains stable.
Popular by topic
Beneficial properties of melatonin
The technique enables a woman to give birth to a healthy child at a late reproductive age
These risk factors have the strongest evidence base. They can be influenced to prevent disease
Checklist of interesting facts about the brain
"Post-viewers" struggle with the consequences of infection for months