SCIENTIFIC EDUCATIONAL CENTER science idea

< Back

Currently, oncogenic mutations in more than 300 genes are known to science, and it often becomes necessary to rank them in order to determine the targets for the main effect of treatment. But in most cases, mutations are ranked by frequency of occurrence, which does not necessarily correspond to their significance in the disease. MIPT scientists proposed to take as a basis the power of mutations. Using bioinformatic analysis of the largest database of oncogenic mutations TCGA PanCanAtlas, they have developed a new way to rank the driver genes that trigger the disease.

Statistically significant representation in the functional categories of the 50 strongest driver genes according to the Driver Strength Index (A) and the Normalized Driver Strength Index (B) — yellow indicates statistical significance

According to the prevailing view today, cancer is initiated and stimulated by mutations in the so-called "driver" genes, while mutations in the "passenger" genes have no effect and simply "travel" with the drivers during somatic evolution. There are several approaches to identifying cancer initiator mutations and ranking them, the main ones are by frequency of occurrence, by effect on the three-dimensional structure of the protein and by effect on the interaction of proteins. These methods have their pros and cons.

One approach is to use the frequency of mutations among patients with a certain type of cancer, usually adjusted for the frequency of background mutations in the gene. This method really allows you to identify the most common mutations, but this does not mean that they are able to cause cancer on their own. At the same time, the mutation may be rare, but sufficient to initiate cancer. The first case will be an example of a common but weak driver, while the second will be a rare but strong driver. Thus, algorithms based on the frequency of mutations cannot reliably determine the strength of driver genes.

There is a large group of algorithms that are aimed at predicting and ranking driver genes according to the effect of mutations on the structure and activity of proteins. These methods can determine whether and to what extent the structure and function of the protein are disrupted, but they are much less suitable for determining the role of a particular protein in the context of other proteins of the cell and its microenvironment. And that is what is crucial for determining whether it will cause cancer. Thus, the degree of influence of a mutation on the structure of a protein does not determine whether this protein is oncogenic.

The third approach determines the effect of mutations on protein interactions by identifying the key molecules with the greatest number of interactions with other proteins in the cell. It is assumed that mutations in such proteins will have the most serious consequences for the cell. This usually turns out to be the case, but much more often it leads to cell death than to oncogenic transformation. Thus, this method is also not suitable for ranking mutations by their strength.

"Previously, we conducted a quantitative assessment of driver mutations and found a very high variability in the number of mutations even among patients with the same type of cancer. Therefore, we asked the question: what is the reason that in some patients one driver mutation is enough, while in others cancer does not develop until dozens of mutations accumulate? We assumed that the main reason is their strength: one strong driver may be equivalent in its tumor-inducing activity to several weak ones.

Therefore, it is statistically more likely that strong drivers are more common in patients with a small number of mutations, because several strong driver mutations are enough to initiate cancer. Conversely, weak drivers are more common in patients with a large number of driver mutations, because weak drivers need a lot to initiate cancer. Based on this principle, we developed mathematical formulas that allowed us to create bioinformatic algorithms for automatically calculating the numerical equivalent of the strength of any driver mutation based on its distribution among patients with different numbers of driver mutations," said the study's lead author Alexey Belikov.

The scientists used formulas for the Driver Strength Index (DSI) and the Normalized Driver Strength Index (NDSI). They conducted a large-scale screening of mutations from the largest TCGA PanCanAtlas database and determined their strength. During the analysis of the data, the scientists found that the strongest drivers often belong to several well-known gene families and make up certain signaling cascades, which are often modified in tumors.

"Our research has shown that the proposed ranking method really reveals a certain biological entity in the driver genes that is not detected by traditional methods, and we believe that this entity is the true power of the driver mutation. Thus, our rating can be used to select the most priority targets for oncotherapy, as well as to select genes for more in-depth research from a scientific point of view. Prioritized genes and signaling pathways probably make the greatest contribution to the occurrence and progression of cancer and may become future therapeutic targets," concluded Sergey Leonov, head of the Laboratory for the Development of Innovative Medicines and Agrobiotechnologies at MIPT, where the study was formulated and conducted.

The article was published in the journal PeerJ Life & Environment
The information is taken from the portal "Scientific Russia" (https://scientificrussia.ru /)

PHOTO: Cancer cells © University of Basel, Biozentrum/Swiss Nanoscience Institute

http://sci-dig.ru