engleski

Microbe Detection Using Deep Learning

Microbes, omnipresent microorganisms invisible to the naked eye, impact many functions in the human body. The ability to detect and classify them is essential in order to discover diseases, prescribe medication, and keep a healthy lifestyle. The goal of this thesis is to develop a method for microbe detection based on a deep learning architecture. The architecture is designed to find suitable representations of signals corresponding to sequenced microbe DNA fragments. After finding the signal repre sentations, an appropriate distance metric is used to separate different species in the latent space. In the end, reads are classified using a suitable classifier.

bioinformatics, deep learning, triplet loss, autoencoder, classification

nije evidentirano