The application of deep learning in miRNA-mRNA target prediction
Great news for the beginning of July!
The nationwide competition for the best master’s thesis on machine learning or data analysis has just announced winners! One of them is Jan Przybyszewski from Sano team.
The research supervisor of the work “The use of deep learning in predicting miRNA-mRNA interactions” is Maciej Malawski, team leader of Extreme-scale Data and Computing in Sano.
What is the application of deep learning in miRNA-mRNA target prediction so about?
Micro RNAs (miRNA) play a key regulatory role in human body, taking part in processes such as cell differentiation, carcinogenesis, or the development of cardiac and gastrointestinal diseases. The ability to deduct, if a given miRNA molecule can target a specific mRNA, can prove to be crucial for our understanding of the aforementioned processes. This work presents a novel target prediction framework, in which the task is considered as a graph classification problem. The evaluation of the proposed graph neural network architecture shows hat the proposed framework is well-suited for target prediction tasks.
MSc thesis was done in collaboration with Sabina Lichołai, MD, from Jagiellonian University Medical College.