Comprehensive anticancer drug response prediction based on a simple cell line-drug complex network model

Abstract

BackgroundAccurate prediction of anticancer drug responses in cell lines is a crucial step to accomplish the precision medicine in oncology. Although many popular computational models have been proposed towards this non-trivial issue, there is still room for improving the prediction performance by combining multiple types of genome-wide molecular data.ResultsWe first demonstrated an observation on the CCLE and GDSC datasets, i.e., genetically similar cell lines always exhibit higher response correlations to structurally related drugs. Based on this observation we built a cell line-drug complex network model, named CDCN model. It captures different contributions of all available cell line-drug responses through cell line similarities and drug similarities. We executed anticancer drug response prediction on CCLE and GDSC independently. The result is significantly superior to that of some existing studies. More importantly, our model could predict the response of new drug to new cell line with considerable performance. We also divided all possible cell lines into “sensitive” and “resistant” groups by their response values to a given drug, the prediction accuracy, sensitivity, specificity and goodness of fit are also very promising.ConclusionCDCN model is a comprehensive tool to predict anticancer drug responses. Compared with existing methods, it is able to provide more satisfactory prediction results with less computational consumption.Electronic supplementary materialThe online version of this article (10.1186/s12859-019-2608-9) contains supplementary material, which is available to authorized users.

Network-based clustering for drug sensitivity prediction in cancer cell lines

“…The novelty of our method is in prior clustering of the data (both cell lines and drugs) by utilizing EMD from the theory of optimal mass transport. It has been shown that similar cell lines by their gene expression profiles exhibit similar responses to the (structurally) similar drugs [18,19]. We improved these methods by focusing on the paired cluster of cell line and drugs to which a new pair of cell line-drug in the test set belongs.…”

“…Increasing the number of decision trees or selected features make the algorithm computationally expensive; however, it does not improve the performance (prediction) significantly. We also compare our result to the network-based model introduced in [19] (CDCN model) which is the extension of Zhang's dual layer integrated cell line-drug network model [18]. The integrated model does not consider the clustering of the data, but it uses the similarity measure between cell lines and drugs via Pearson correlation.…”

“…Following this idea, we show that prior clustering of cell lines and drugs significantly improves the prediction. We compare our results with the Zhang et al extended method in [19] to show that random forest is more effective than the prediction formula provided by [18,19]. Moreover, in both of these network-based models, the Pearson correlation coefficent has been incorporated to measure the similarity of the samples.…”

“…The decay parameter ζ = (α,β) can be optimized on the training set Γ by minimizing the error of response prediction as follows [19]:…”

Quantitative Structure-Mutation-Activity Relationship Tests (QSMART) Model for Protein Kinase Inhibitor Response Prediction

“…11 Chang et al [33] CNN (RF, SVM) SNP 5% leave-out R 2 = 0.843 2018.07.01 Cichonska et al [31] Kernel SNP, MET, EXP, CNV 10-fold CV r = 0.858 2018.09.14 Le et al [19] Network (Kernel) MUT, EXP 5-fold CV r = 0.804 2018.09.14 Juan-Blanco et al [20] Network MUT, EXP LOOCV AUC = ∼0.72 2018.10.10 Yang et al [21] Network + SVM (Kernel) MUT, MET, CNV, PPI 5-fold CV AUC = 0.788 2018.12.07 Liu et al [22] Network EXP 10-fold CV r = 0.73 2019.01. 22 Wei et al [23] Network EXP LOOCV r = 0.63 2019.01. 31 Wang et al [15] EN EXP, PWY 10-fold CV MSE = ∼2.8 2019.01.…”