Skip to main content
. 2023 Apr 19;7:38. doi: 10.1038/s41698-023-00375-y

Table 2.

Model details and validation results for the prediction of cisplatin response using CisSig in GDSC dataset.

Input Output Method Included data Metric Value
CisSig Score IC50 (continuous) Simple Linear Regression All Corr. Coef. 0.51
CisSig Score IC50 (continuous) Simple Linear Regression Quintiles Corr. Coef. 0.74
All gene expression IC50 (continuous) Elastic Net Linear Regression All Corr. Coef. 0.63
All gene expression IC50 (continuous) Elastic Net Linear Regression Quintiles Corr. Coef. 0.79
All gene expression IC50 (continuous) L1 Linear Regression All Corr. Coef. 0.63
All gene expression IC50 (continuous) L1 Linear Regression Quintiles Corr. Coef. 0.79
All gene expression IC50 (continuous) L2 Linear Regression All Corr. Coef. 0.63
All gene expression IC50 (continuous) L2 Linear Regression Quintiles Corr. Coef. 0.81
All gene expression IC50 (binary) Simple Logistic Regression All AUC 0.79
All gene expression IC50 (binary) Simple Logistic Regression Quintiles AUC 0.90
All gene expression IC50 (binary) Elastic Net Logistic Regression All AUC 0.82
All gene expression IC50 (binary) Elastic Net Logistic Regression Quintiles AUC 0.94
All gene expression IC50 (binary) L1 Logistic Regression All AUC 0.82
All gene expression IC50 (binary) L1 Logistic Regression Quintiles AUC 0.94
All gene expression IC50 (binary) L2 Logistic Regression All AUC 0.81
All gene expression IC50 (binary) L2 Logistic Regression Quintiles AUC 0.95
All gene expression IC50 (binary) SVM (linear kernel) All AUC 0.82
All gene expression IC50 (binary) SVM (linear kernel) Quintiles AUC 0.93
All gene expression IC50 (binary) SVM (polynomial kernel) All AUC 0.78
All gene expression IC50 (binary) SVM (polynomial kernel) Quintiles AUC 0.94
All gene expression IC50 (binary) Random Forest All AUC 0.81
All gene expression IC50 (binary) Random Forest Quintiles AUC 0.91