Search for genomic associations
Sergey Nuzhdin, Post-science
As part of a joint project of Post-Science and Peter the Great St. Petersburg Polytechnic University, we are publishing a text by Sergey Nuzhdin, Candidate of Biological Sciences, dedicated to the research of associations of genes and mutations of humans and simple organisms.
There are genes and mutations in the human genome that are responsible for the fact that people develop, look and think differently. A statistical search mechanism is used to detect associations of these genes. This area of research is now called precision medicine. Using statistical methods, it is possible to estimate how strongly the genes of different people are associated. For example, you can observe a group of people with high blood pressure and find a mutation that is present in the whole group. Today it is important to look for genomic associations, because two thousand mutations of different genes may be responsible for the same disease. If scientists understand exactly how these mutations lead to the development of diseases, they can be classified in such a way as to correct the process that leads to the disease. There are many mutations and few processes, so fewer drugs may be needed to treat diseases.
But the mechanism of searching for such associations is not yet fully clear. Currently, it is easiest to understand this by experimenting on flies, since it is easier to learn from simple organisms. Scientists have not yet learned to look for associations in humans, because the human genome is large, its processes are still poorly understood. The search for genomic associations and the analysis of genomic chains in flies allows us to predict how these flies will look when various genes break down.
Today, these studies are carried out on flies in order to develop a certain search model and then apply it to clarify genomic associations in humans. The study of genomic mutations is necessary in order to find out how to prevent certain genetic diseases or reverse the process of their development. These studies stand at the intersection of population genetics and systems biology. The methods of searching for genomic associations in humans and simple organisms are the same. But scientists start their research with simple organisms, because there are 3 billion nucleotides in the human genome, and, for example, flies have 150 million. Therefore, it is cheaper and faster to experiment on flies.
Genomic associations
Genomic associations were first studied in the 1980s by scientists Trudy McKay from North Carolina and Chuck Langley from the University of California. At that time, scientists had not yet identified the complete human genome. McKay and Langley studied mutations resulting from the mixing of mobile elements – viruses that live in the human genome and spread through it. McKay and Langley observed when mutations occur in these mobile elements of flies and whether these effects can be seen by phenotype. Scientists began observations with such an abstract feature as the number of bristles. Bristles in flies are formed as a result of the fact that cells close to each other put pressure on the formation of similar bristles in other cells. This mechanism is called lateral inhibition. The delta and notch genes are responsible for this. Humans have the same genes, and they are also responsible for similar processes. Strong expression of these genes causes the growth of bristles at a greater distance. Weak expression causes the growth of bristles on the nearest cells. McKay and Langley found that when mobile elements are embedded in these genes, the bristles grow denser. This was the first article devoted to genomic associations. It was published in 1981 in the journal Nature. In the next 10 years, a boom in such works began.
After scientists identified the human genome, they began to look for genome mutations. There was a large wave of research in which scientists tried to find genes that explain schizophrenia, body size, IQ and much more. Whole countries in the world began to engage in such research. For example, it is very convenient to conduct such studies in Iceland, because Icelanders descended from a very small number of ancestors and the population expanded rapidly. Therefore, mutations that appeared in a small number of ancestors are now present in a large number of people. This makes it possible to obtain statistically correct predictions. Now, in order to conduct such studies, samples from 30 to 150 thousand people are sequenced or genotyped. For example, out of 150 thousand people, 10 thousand had the same mutation. And it turns out that all of these 10 thousand, for example, have a predisposition to schizophrenia. Such studies have been called "genome-wide association search", or genome-wide association.
We know that some people get cancer and others don't. There are genetic bases for this. One of the assumptions is that one person has a higher mutation rate than another. Therefore, mutations of one person can lead to cancer, and the other – not. It depends on the frequency of mutations. In order to study the possibility of cancer, it is possible to check the frequency of mutations in a group of people using genome-wide methods. Let's say we observe 3 thousand people with cancer and 2 thousand people without cancer and check whether there are significant differences between them at the level of mutations not in cancer, but in normal tissues. If mutations occur more often in normal tissues in those people who have cancer, this may be due to the fact that their mutation frequency is higher.
Further research
It is expected from genomic association studies that we will soon be able to predict mutations and the risk of diseases. To date, such predictions give about 10% accuracy. For example, if we want to predict a person's height, we can measure the height of 5 thousand people, sequence their genomes and find a thousand mutations that are associated with a difference in height. We will get a 10% probability. And if you ask your parents what their height is, and, based on this, make an assumption about their height, the probability of such a prediction will be 60%. This is called heredity. Therefore, it is not very clear: to predict diseases based on the diseases of relatives or from genomic information. So now we are developing new methods. The next stage of research is moving away from complete genomic associations and statistical methods. We are trying to understand which mutations break certain mechanisms and how these mechanisms predict phenotypes.
About the author:
Sergey Nuzhdin – Candidate of Biological Sciences, Professor, Scientific Supervisor of the Research Institute "Systems Biology and Bioinformatics" of Peter the Great SPbPU, Professor, Biology, College of Letters, Arts, and Sciences, University of Southern California.
Portal "Eternal youth" http://vechnayamolodost.ru
26.05.2016