Canada Spends Logo

Explore our Budget 2026 analysis

Grant: Master's Award: Canada Graduate Scholarships

McMaster University

$17,500.00

Leveraging machine learning for the discovery of structurally novel antibiotics against multidrug-resistant bacterial pathogens

FY 2023-24

Summary
The global spread of antibiotic resistance has significantly diminished available treatment options for bacterial infections. The ESKAPE pathogens are a group of multi-drug resistant bacteria that are now resistant to last resort antibiotics. Alarmingly, no novel antibiotic classes have emerged for clinical use since 1987, attributed to bottlenecks in conventional antibiotic discovery pipelines. Fortunately, due to rapid increases in computational power and advancements in artificial intelligence, we can leverage machine learning (ML) algorithms to virtually screen vast chemical libraries and prioritize molecules for testing in the lab. This increases the rate at which we can discover antibiotic candidates in the early stages of preclinical research. My project is to leverage an ensemble of ML models to predict antibacterial activity against each of the ESKAPE pathogens and facilitate the discovery of promising antibiotic candidates. I tested a couple thousand molecules against each pathogen and used this data to train ML models to learn the structural features of molecules that are antibacterial. I then employed my models to predict on over 12 million molecules from various chemical databases, and used these predictions to prioritize specific molecules to test in the lab. To date, I have tested one molecule that I found to be antibacterial against five of the pathogens, with a potency comparable to that of currently used antibiotics. As I continue to test more molecules, I will select the most promising ones to investigate the mechanism of action and assess toxicity and infection clearance in mouse models. Importantly, the ESKAPE Model will be released as a free online resource where users can input molecules and receive predictions across all models.
Program
Master's Award: Canada Graduate Scholarships
Location
Ontario, Canada
Recipient Type
Academia
Type
Grant
Source
CIHR Funding Decisions Database

As published

Id
505855
Name
Arnold Patricia A
Region
Ontario
Country
Canada
Orgtype
Institution Paid; Research Institution
Prcname2
Special Cases - Awards Programs
Projecttitle
Leveraging machine learning for the discovery of structurally novel antibiotics against multidrug-resistant bacterial pathogens
Programtype2
Studentship Programs
Orgnameinp2
McMaster University
Cihramount2
$17,500
Cihrequipment2
$0
Programname2
Master's Award: Canada Graduate Scholarships
Keyworddelim
Antibiotic Discovery; Antibiotic Resistance; Antibiotics; Artificial Intelligence; Bacteria; Chemiformatics; Deep Learning; Growth Inhibition Screening; Infectious Disease; Machine Learning
Abstract
The global spread of antibiotic resistance has significantly diminished available treatment options for bacterial infections. The ESKAPE pathogens are a group of multi-drug resistant bacteria that are now resistant to last resort antibiotics. Alarmingly, no novel antibiotic classes have emerged for clinical use since 1987, attributed to bottlenecks in conventional antibiotic discovery pipelines. Fortunately, due to rapid increases in computational power and advancements in artificial intelligence, we can leverage machine learning (ML) algorithms to virtually screen vast chemical libraries and prioritize molecules for testing in the lab. This increases the rate at which we can discover antibiotic candidates in the early stages of preclinical research. My project is to leverage an ensemble of ML models to predict antibacterial activity against each of the ESKAPE pathogens and facilitate the discovery of promising antibiotic candidates. I tested a couple thousand molecules against each pathogen and used this data to train ML models to learn the structural features of molecules that are antibacterial. I then employed my models to predict on over 12 million molecules from various chemical databases, and used these predictions to prioritize specific molecules to test in the lab. To date, I have tested one molecule that I found to be antibacterial against five of the pathogens, with a potency comparable to that of currently used antibiotics. As I continue to test more molecules, I will select the most promising ones to investigate the mechanism of action and assess toxicity and infection clearance in mouse models. Importantly, the ESKAPE Model will be released as a free online resource where users can input molecules and receive predictions across all models.
Competitiondate
202312
Approvedterm2
1 yr 0 mth
Pinamesdelim
Arnold, Patricia A
Approvedterm
1 an 0 mois
Cihrcontribution2
~ $17,500~ $0~