Department of Computer Science: MSc Thesis Presentation
Date: Wednesday, 15 June 2022
Deep Learning Methodologies in Drug Kinase prediction
Author: Severi Vapalahti
Supervisor: Juho Rousu
Advisor: Tianduanyi Wang
Abstract: The misbehaviour of the enzymatic protein kinases can lead to development of tumours. Small molecule kinase inhibitors are known to be effective therapeutics in cancer treatment, but it is difficult to find selective drugs. To avoid laborious biochemical experiments, the drug-target space can be scanned with computer models. Multiple predictive algorithms have been developed based on different machine learning paradigms. This thesis introduces a predictive recurrent convolutional neural network regression model inspired by two previous deep learning algorithms. The model had good performance when tested with a validation set isolated from the training data. However, some of its performance reduced when its generalizability to different dataset was tried out. Nevertheless, taking into account the many potential ways the model can be improved and comparatively light design, similar architectures can have prospect for scanning potential hits in early stage of discovering therapeutic kinase inhibitors.