The analysis of DNA and genomic data, as well as the extraction of variants or mutations, is very complex. Automatic detection of any of them presenting a potential risk (either because they have a high probability of contributing to the development of diseases such as cancer, or because they present unusual characteristics) and therefore appropriate to analyze them in depth, is of great interest.
With this in mind, researchers from the Group for Visual Communication Applications (GATV) of the Polytechnic University of Madrid (UPM) in Spain have developed a system that provides, through artificial intelligence, the probability that the variables under study may be malignant (and therefore of interest to the analysis) or may be benign variants. within the human genome. The work developed and the proposed tool could help study genetic alterations in human cancers that allow targeted therapies for precise oncology based on next-generation sequencing of personal data.
These researchers from UPM’s GATV conducted a study using several traditional, and other new, techniques of automatic learning (an artificial intelligence method) to classify somatic mutations.
As a result, they have developed a classification tool that aggregates a large number of known variants, both malignant and benign, collected in approved clinical studies and made available to the general public in open databases.
A symbolic artistic restatement of the concept of artificial intelligence, represented here by the binary code of a computer program, used in the fight against cancer, represented here by the cancer cell. (Illustration: Jorge Monchi for NCYT from Amazings)
To learn more about each of these mutations, they used an annotation software tool called ANNOVAR, which aggregates databases and other algorithms provide additional details about the mutations. 70 annotations (that is, 70 numerical values describing each variant collected) were used as inputs to different AI models aiming to obtain the probability that each of these mutations is benign or malignant.
The same dataset was tested on existing classifiers that also used AI models, among other proposals, to compare the effectiveness of the detectors. After comparing solutions using known machine learning rating metrics, it can be concluded that the implications of these results are highly relevant, as they show that this proposed rating tool outperforms the rest of the studied ones.
The results obtained were very promising, as the best AI models were able to correctly classify about 80% of the potentially dangerous mutations. As researcher Anaida Fernández García who was part of the work notes, “From a medical perspective, these tools are very useful when doing a genetic analysis of a patient, since the percentage of potentially malignant variants is usually very high.” Small compared to benign, so the The ability to find them quickly and automatically presupposes a significant reduction in the man-hours that could be devoted to the direct study of these potential discoveries.”
All of this work was done thanks to the European project Genomed4All led by UPM.
The new study is titled “DrOGA: An Artificial Intelligence Solution for Driver Status Prediction of Genome Mutations.” It has been published in the academic journal Precision Cancer Medicine. (Source: UPM)
“Creator. Devoted pop culture specialist. Certified web fanatic. Unapologetic coffee lover.”