Identifying and analysing significantly impacted pathways: a complex review and assessment

Abstract

The bachelor's qualification work was completed by a student of the KN-414 group Boychuk Oleslav-Ivan Mychailovych. The topic is "Identifying and analysing significantly impacted pathways: a complex review and assessment". The work is aimed at obtaining a bachelor's degree in 122 "Computer Science". The object of research is the intersection between bioinformatics and computer science, especially deep learning neural networks. The subject of research are dataset of gene expression data, which was collected from combining two projects: the Cancer Genome Atlas Program and The Genotype Tissue Expression. The research methods are a complex system for learning a neural network and extracting useful features from it using an layer wise relevance propagation. These useful features will be used in further research to identify and analyse significantly impacted pathways. The goal of the research is to increase the accuracy of solving the detect and classification problem of cancer diseases. With the help of modern methods of classification by neural networks, namely the determination of significant genes and pathways.