Drug Sensitivity Prediction From Cell Line-Based Pharmacogenomics Data: Guidelines for Developing Machine Learning Models
Hossein Sharifi Noghabi1, Petr Smirnov2, Soheil Jahangiri-Tazehkand2, Anthony Mammoliti2, Sisira Kadambat Nair3, Casey Hon2, Arvind Mehr3, Martin Ester 1, Benjamin Haibe-Kains2 1 Simon Fraser University 2 University of Toronto 3 University Health Network
Environment
R
Ubuntu
Packages:
pytorch, torchvision, cpuonly, pandas, matplotlib
Code
R, pytorch
Data
Pan-cancer datasets use are the Cancer Therapeutics Response Portal (CTRPv2) , the Genentech Cell Line Screening Initiative (gCSI) and the Genomics of Drug Sensitivity in Cancer (GDSCv1 and GDSCv2)
Results
What's inside
The authors develop a set of guidelines for different aspects of training gene expression-based predictors using cell line datasets which challenge many current practices, e.g. choice of training datasets.
Who uses it
Cancer researchers focused on drug sensitivity prediction for precison medicine.
Why we like this
Application of these guidelines in future studies will enable the development of more robust preclinical biomarkers.
Copy this HTML code to embed this Capsule
<script src="https://codeocean.com/widget.js?slug=7358839" async></script>
For any questions about the Compute Capsule, please contact us at [email protected].
