aalto1 untyped-item.component.html
Protein language models (pLMs): Utility and limitations for predicting material-related properties of proteins
Loading...
URL
Journal Title
Journal ISSN
Volume Title
Kemian tekniikan korkeakoulu |
Bachelor's thesis
Unless otherwise stated, all rights belong to the author. You may download, display and print this publication for Your own personal use. Commercial use is prohibited.
Authors
Date
Department
Major/Subject
Mcode
CHEM3054
Degree programme
Language
en
Pages
24
Series
Abstract
The recent integration of deep learning with protein science has given rise to protein language models (pLMs), which adapt natural language processing techniques to analyze and predict protein sequences. This thesis explores the application of pLMs in predicting material-related properties of fibrous proteins—key components in biomaterials such as elastin, spider silk, and mussel byssus core. These fibrous proteins exhibit highly repetitive sequences and distinct structure-function relationships, making them ideal candidates for computational modeling. The study compares classical machine learning methods relying on evolutionary information with modern pLM-based approaches, evaluating their effectiveness in capturing structural motifs, phase behavior, and mechanical properties critical to biofabrication. While pLMs offer advantages in scalability, efficiency, and applicability to orphan proteins, they face limitations related to sequence length, data bias, and training distribution gaps. By assessing both the strengths and constraints of pLMs, this work highlights their potential to further the development of protein-based functional materials, contributing to sustainability, biotechnology, and materials engineering. Future research should prioritize improving model generalizability and interpretability to fully leverage pLMs for protein design tasks.