Title: Researchers Develop AI System for Detecting Potentially Malignant DNA Mutations
Subtitle: Artificial intelligence tool aids precision oncology based on genomic data analysis
The analysis of DNA and genomic data, along with the identification of variants or mutations, is a complex task. Detecting potentially risky mutations, which may contribute to the development of diseases like cancer or possess unusual characteristics, is crucial for in-depth analysis. In response to this need, researchers from the Visual Telecommunications Application Group (GATV) at the Polytechnic University of Madrid (UPM) have developed an artificial intelligence system that determines the probability of the studied variants being potentially malignant or benign within the human genome.
The team at GATV conducted a study using traditional techniques as well as artificial learning methods to classify somatic mutations. They developed a classification tool that incorporates a vast collection of known variants, both malignant and benign, sourced from endorsed clinical studies and available open databases.
To understand each mutation in detail, the researchers employed ANNOVAR, an annotation software tool that integrates various databases and algorithms. By utilizing 70 annotations that define each variant, artificial intelligence models were employed to compute the probability of each mutation being benign or malignant.
The effectiveness of the proposed classification tool was benchmarked against existing classifiers that also utilized artificial intelligence models. The results showed that the newly developed tool surpassed other existing methods, demonstrating its superiority in classification accuracy. The implications of these findings are significant as they provide a more efficient approach to detecting potentially dangerous mutations.
The best artificial intelligence models in this study were able to accurately classify approximately 80% of potentially dangerous mutations. Anaida Fernández García, a researcher on the team, highlighted the practical benefits of such tools in genomic analysis, stating that “the percentage of potentially malignant variants is usually very small compared to benign ones, so being able to quickly and automatically identify them significantly reduces the working hours required for their study.”
This research was made possible by the European project Genomed4All, led by the UPM. The study, titled “DrOGA: An Artificial Intelligence Solution for Driver-Status Prediction of Genomics Mutations,” has been published in the academic journal Precision Cancer Medicine.
The development of this AI system presents a vital advancement in precision oncology, offering a faster and more accurate approach to the analysis of genomic alterations in human cancers. The ability to identify potentially malignant mutations promptly can lead to targeted therapies based on next-generation sequencing personal data.