Machine Learning Prediction of Cancer Cell Sensitivity to Drugs Based on Genomic and Chemical Properties
Problem
Predicting how sensitive specific cancer cell lines are to drug treatment, a step toward personalised cancer therapy.
Why it matters
Personalised oncology depends on predicting which drugs will work against a given tumour.
What the AI did
Machine learning models integrated genomic features of cell lines with chemical properties of drugs to predict IC50 drug response values.
Results
R² of 0.72 in 8-fold cross-validation and 0.64 on an independent blind test for IC50 prediction.
Impact
Demonstrated that combined genomic and chemical features predict cell-line drug response with moderate accuracy.
Record
| Venue | PLoS ONE |
|---|---|
| Published | 2013 |
| Work type | article |
| Authors | Michael P. Menden, Francesco Iorio, Mathew J. Garnett, Ultan McDermott, Cyril H. Benes, Pedro J. Ballester, Julio Sáez-Rodríguez |
| Citations | 543 (OpenAlex, at import time) |
| References | 40 |
| Model or method | not stated in source |
| Organization | not stated in source |
| Primary source | https://doi.org/10.1371/journal.pone.0061318 |
| Open access | gold |
| Keywords | machine learning, cancer, drug sensitivity, IC50, genomics, personalised medicine |
Importance rating
significant — thousands to millions affected
Abstract states predicting cancer therapy response is a major goal of modern oncology that should lead to personalised treatment; affects many patients, though no scale numbers are given.