Supervisor: Mgr. Tomáš Vinař, PhD.
MinION is a sequencing platform that produces sequences of electrical signals
from DNA molecules. These signals can then be translated into a string
representation of the DNA (a string over alphabet {A,C,G,T}). Such translation
is non-trivial and requires advanced machine learning methods, such as
recurrent neural networks.
This string representation does not cover various base modifications, however
such modification cause changes in the electrical signals produced by the
device. The goal of this thesis is to develop a method to identify these
modifications by unsupervised (e.g. outlier detection) or semi-supervised
methods.