Canada Spends Logo

Explore our Budget 2026 analysis

Grant: Master's Award: Canada Graduate Scholarships

University of British Columbia

$17,500.00

Machine-Learning Aided Discovery of targeted ER-AF2 PROTACs

FY 2021-22

Summary
Proteolysis targeting chimera (PROTAC) aims at selectively degrading biological targets of interest (TOI) intracellularly using host cell's ubiquitin ligases thereby achieving lasting ablation of activity. PROTACs carry special interest in this project since the objective is to target conventionally 'undruggable' targets - estrogen receptor (ER) - through unconventional allosteric sites - co-activator binding domain (AF2) of ER - which is a relatively solvent-exposed pocket. PROTACs consist of three chemical moieties: 1) a binder for the to-be-degraded TOI. 2) a binder to the host cell's E3 ubiquitin ligases. 3) a linker that connects parts 1) and 2). Finding the best combination of those three elements to form a stable ternary complex constitutes the challenge in the design of new PROTACs. First, formation of the ternary complex is thermodynamically driven by cooperativity of favorable interactions between the TOI and the E3 ligase. Second, the choice of the E3 ligase is predicated on the abundance of the E3 ligase concentration in target tissues. Third, the linker needs to be minimal in length in order to limit off target effects and ideally participate in promoting the ternary complex. In order to predict choice of E3 ligase and optimal linker type and length, we will employ a machine learning model that will train on all available empirical active/inactive PROTAC data points (more than 1200 PROTACs were tested to date) and incorporate protein surface-surface encodings from protein-protein dockings in order to derive TOI specific predictions. Additionally, tissue-specific expression profiles will be curated. The application of this project may generalize beyond PROTACs to lysosomal targeting chimeras, molecular glues, and other novel non-inhibition mediated schemes.
Program
Master's Award: Canada Graduate Scholarships
Location
British Columbia, Canada
Recipient Type
Academia
Type
Grant
Source
CIHR Funding Decisions Database

As published

Id
467045
Name
Mslati Hazem
Region
British Columbia
Country
Canada
Orgtype
Institution Paid; Research Institution
Prcname2
Special Cases - Awards Programs
Projecttitle
Machine-Learning Aided Discovery of targeted ER-AF2 PROTACs
Programtype2
Studentship Programs
Orgnameinp2
University of British Columbia
Cihramount2
$17,500
Cihrequipment2
$0
Programname2
Master's Award: Canada Graduate Scholarships
Keyworddelim
Af2; Breast Cancer; Coactivator Domain; Estrogen Receptor; Machine Learning; Modelling Protacs; Predicting Active Protacs; Protac; Protein Degradation; Protein-Protein Docking
Abstract
Proteolysis targeting chimera (PROTAC) aims at selectively degrading biological targets of interest (TOI) intracellularly using host cell's ubiquitin ligases thereby achieving lasting ablation of activity. PROTACs carry special interest in this project since the objective is to target conventionally 'undruggable' targets - estrogen receptor (ER) - through unconventional allosteric sites - co-activator binding domain (AF2) of ER - which is a relatively solvent-exposed pocket. PROTACs consist of three chemical moieties: 1) a binder for the to-be-degraded TOI. 2) a binder to the host cell's E3 ubiquitin ligases. 3) a linker that connects parts 1) and 2). Finding the best combination of those three elements to form a stable ternary complex constitutes the challenge in the design of new PROTACs. First, formation of the ternary complex is thermodynamically driven by cooperativity of favorable interactions between the TOI and the E3 ligase. Second, the choice of the E3 ligase is predicated on the abundance of the E3 ligase concentration in target tissues. Third, the linker needs to be minimal in length in order to limit off target effects and ideally participate in promoting the ternary complex. In order to predict choice of E3 ligase and optimal linker type and length, we will employ a machine learning model that will train on all available empirical active/inactive PROTAC data points (more than 1200 PROTACs were tested to date) and incorporate protein surface-surface encodings from protein-protein dockings in order to derive TOI specific predictions. Additionally, tissue-specific expression profiles will be curated. The application of this project may generalize beyond PROTACs to lysosomal targeting chimeras, molecular glues, and other novel non-inhibition mediated schemes.
Competitiondate
202112
Approvedterm2
1 yr 0 mth
Pinamesdelim
Mslati, Hazem
Approvedterm
1 an 0 mois
Cihrcontribution2
~ $17,500~ $0~