Detection of Modified Nucleotide Clusters in Nanopore Sequenced RNA Reads

The main goal of this thesis was to detect groups of modified nucleotides in RNA reads obtained by third generation sequencers. Reads obtained by third-generation sequencers often contain a large amount of noise. In this work, detailed analyzes of the data were performed in order to try to remove the noise and so that the data could be better understood. The related work achieved in the area of this problem is shown. Clustering was performed by various methods such as K-means, Graph Spectral Clustering and Non-negative matrix factorization. Clustering methods were tested on simulated datasets as well as on real datasets. The results are presented in tables and graphs. The source code of the implemented methods can be found at: