Iterative sure independent ranking and screening for drug response prediction
- PDF / 1,196,783 Bytes
- 9 Pages / 595.276 x 790.866 pts Page_size
- 118 Downloads / 183 Views
RESEARCH
Open Access
Iterative sure independent ranking and screening for drug response prediction Biao An1†, Qianwen Zhang2,3†, Yun Fang1, Ming Chen2,3* and Yufang Qin2,3* From 15th International Symposium on Bioinformatics Research and Applications (ISBRA) Barcelona, Spain. 3-6 June 2019
Abstract Background: Prediction of drug response based on multi-omics data is a crucial task in the research of personalized cancer therapy. Results: We proposed an iterative sure independent ranking and screening (ISIRS) scheme to select drug responseassociated features and applied it to the Cancer Cell Line Encyclopedia (CCLE) dataset. For each drug in CCLE, we incorporated multi-omics data including copy number alterations, mutation and gene expression and selected up to 50 features using ISIRS. Then a linear regression model based on the selected features was exploited to predict the drug response. Cross validation test shows that our prediction accuracies are higher than existing methods for most drugs. Conclusions: Our study indicates that the features selected by the marginal utility measure, which measures the conditional probability of drug responses given the feature, are helpful for drug response prediction. Keywords: SIRS, Drug response, ISIRS, CCLE
Background A major goal in cancer research is to select an efficacious drug or drug combinations for each individual patient based on their genomic and transcriptomic profiles [1]. To get a much more comprehensive understanding of the potential genetic makeup of a patient, researchers have tried multiomics data including protein concentration, gene expression and genetic mutations. However, the methodology of translating the genetic measurements to predictive models for assisting therapeutic decisions is still a challenge. Researchers have tried many methods to find biomarkers and predict drug sensitivity. These methods are mainly based on gene expression measurements. * Correspondence: [email protected]; [email protected] † Biao An and Qianwen Zhang contributed equally to this work. 2 College of Information Technology, Shanghai Ocean University, Shanghai, China Full list of author information is available at the end of the article
Staunton et al. proposed a weighted voting classification strategy to classify each cell line as sensitive or resistant for each drug based on the NCI-60 gene expression data [2]. Riddick et al. developed a novel multistep regression model for drug response using Random Forest [3]. However, the biomarker of a certain drug for different cancer types may be different because of the heterogeneity of different cancers, so it is more realistic to focus on some specific type of cancers. Lee et al. developed a genetic algorithm termed as “coexpression extrapolation”, which can accurately predict drug sensitivity of bladder cancer cell lines and clinical responses of breast cancer patients treated by commonly used chemotherapeutic drugs [4]. Holleman et al. used 14,500 probe sets to identify differentially expressed genes in drug-sensitive and drug
Data Loading...
