AI Enabled Drug Design and Side Effect Prediction Powered by Multi-Objective Evolutionary Algorithms & Transformer Models

Due to the large search space and conflicting objectives, drug design and discovery is a difficult problem for which new machine learning (ML) approaches are required. Here, the problem is to invent a method by which new, therapeutically useful, compounds can be discovered; and to simultaneously avoid compounds which will fail clinical trials or pass unwanted effects onto the end patient. By extending current technologies as well as adding new ones, more design criteria can be included, and more promising novel drugs can be discovered. This work advances the field of computational drug design by (1) developing MOEA-DT, a non-deep learning application for multi-objective molecular optimization, which generates new molecules with high performance in a variety of design criteria; and (2) developing SEMTL-BERT, a side effect prediction algorithm which leverages the latest ML techniques and datasets to accomplish its task. Experiments performed show that MOEA-DT either matches or outperforms other similar methods, and that SEMTL-BERT can enhance predictive ability.

Name thesis:degree_name

M.Sc. Computer Science

Level thesis:degree_level

Masters

Discipline thesis:degree_discipline

Faculty of Mathematics and Science

Department dc:contributor.department

Department of Computer Science

Grantor

Brock University

Year dc:date.issued

2023

Author dc:creator

• Grantham, Karl

• Artificial Intelligence, Machine Learning, Drug Design and discovery

Statement dc:rights

• Attribution-NonCommercial-ShareAlike 4.0 International

Licence dc:rights.uri

• http://creativecommons.org/licenses/by-nc-sa/4.0/

Language dc:language.iso

eng

Handle dc:identifier.uri

http://hdl.handle.net/10464/18160

OAI identifier oai:identifier

oai:brocku.scholaris.ca:10464/18160