Date of Publication

2023

Document Type

Master's Thesis

Degree Name

Master of Science in Computer Science

Subject Categories

Software Engineering

College

College of Computer Studies

Department/Unit

Software Technology

Thesis Advisor

Anish M.S. Shrestha

Defense Panel Chair

Charibeth Cheng

Defense Panel Member

Llewelyn Espiritu
Anish M.S. Shrestha

Abstract (English)

With the rise of antimicrobial resistance that decreases the effectiveness of antibiotics in treating bacterial infections, phage therapy is being studied as an alternative to antibiotics. Phage therapy is the use of phages to treat bacterial infections by letting the phages infect and lyse the bacterial pathogen at the site of infection. Phages are known to be able to infect a narrow range of hosts only, but laboratory experiments to verify an interaction between a phage and a bacterium are both costly and time-consuming. To mitigate this, several studies explored the use of machine learning classifiers to predict whether a phage host interacts or not. In this study, we explored different kinds of protein representation, including protein embeddings that are generated by protein language models, that can serve as an input to machine learning classifiers. Our experiments show that protein embeddings do not necessarily improve classifier performance compared to using the conventional k-mer profile representation.

Abstract Format

html

Abstract (Filipino)

-

Abstract Format

html

Language

English

Keywords

Bacteriophages; Protein engineering; Proteins -- Data processing

Upload Full Text

wf_yes

Embargo Period

12-10-2023

Download

Share

COinS