Canada Spends Logo

Explorez notre analyse du budget 2026

Grant: Master's Award: Canada Graduate Scholarships

University of Toronto

$17,500.00

Predicting Cancer Drug Response with Deep Learning

EF 2017-18

Résumé
Cancer treatment outcomes can be improved by treating patients with the optimal drugs from the time they are diagnosed. This is a main goal of precision medicine, yet success in identifying the best personalized treatment for a given patient is limited to cancer subtypes that have been well-studied and for which there are known mutations. There is a limited amount of clinical trial data that can be used to associate genetic traits of patients with their response to available drugs, so analyzing this data on its own is not enough. However, there are studies that have been performed on cancer cell lines in vitro which generated significant amounts of data, but these data have their limitations since they are approximations of the human body. My research focuses on building predictive models of drug response using deep learning in the in vitro setting. In order for this to be effective, I will use multiple types of data in addition to gene expression such as mutation status and DNA methylation. The models I will design will incorporate prior biological knowledge regarding the relationships between genes, as well as the relationships between the different types of data mentioned previously. Furthermore, the models will be designed in such a way that the confidence they report will correspond to the likelihood of a treatment being effective, thereby allowing the ranking of drugs based on predicted efficacy. If demonstrated to be better than the existing drug assignment methods, patients can be saved from undergoing needlessly harmful treatments that likely won't be effective for them, and the healthcare system as a whole will be more efficient.
Programme
Master's Award: Canada Graduate Scholarships
Bénéficiaire
Emplacement
Ontario, Canada
Type de bénéficiaire
Academia
Type
Grant
Source
CIHR Funding Decisions Database

Tel que publié

Id
386568
Name
Adam George A
Region
Ontario
Country
Canada
Orgtype
Institution Paid; Research Institution
Prcname2
Special Cases - Awards Programs
Projecttitle
Predicting Cancer Drug Response with Deep Learning
Programtype2
Studentship Programs
Orgnameinp2
University of Toronto
Cihramount2
$17,500
Cihrequipment2
$0
Programname2
Master's Award: Canada Graduate Scholarships
Keyworddelim
Biological Networks; Deep Learning; Drug Response Prediction; Pharmacogenomics; Representation Learning; Structured Neural Networks
Abstract
Cancer treatment outcomes can be improved by treating patients with the optimal drugs from the time they are diagnosed. This is a main goal of precision medicine, yet success in identifying the best personalized treatment for a given patient is limited to cancer subtypes that have been well-studied and for which there are known mutations. There is a limited amount of clinical trial data that can be used to associate genetic traits of patients with their response to available drugs, so analyzing this data on its own is not enough. However, there are studies that have been performed on cancer cell lines in vitro which generated significant amounts of data, but these data have their limitations since they are approximations of the human body. My research focuses on building predictive models of drug response using deep learning in the in vitro setting. In order for this to be effective, I will use multiple types of data in addition to gene expression such as mutation status and DNA methylation. The models I will design will incorporate prior biological knowledge regarding the relationships between genes, as well as the relationships between the different types of data mentioned previously. Furthermore, the models will be designed in such a way that the confidence they report will correspond to the likelihood of a treatment being effective, thereby allowing the ranking of drugs based on predicted efficacy. If demonstrated to be better than the existing drug assignment methods, patients can be saved from undergoing needlessly harmful treatments that likely won't be effective for them, and the healthcare system as a whole will be more efficient.
Competitiondate
201712
Approvedterm2
1 yr 0 mth
Pinamesdelim
Adam, George A
Approvedterm
1 an 0 mois
Cihrcontribution2
~ $17,500~ $0~