NeuPD-A Neural Network-Based Approach to Predict Antineoplastic Drug Response

Area of Science:

• Bioinformatics
• Computational Biology
• Genomics

Background:

• Personalized medicine relies on accurate in silico drug response analysis.
• Existing anti-cancer drug sensitivity prediction methods require improvement.

Purpose of the Study:

• To propose and validate the NeuPD framework for predicting anti-cancer drug efficacy.
• To integrate cell line genomic features and drug fingerprints for enhanced prediction.

Main Methods:

• Utilized Genomics of Drug Sensitivity in Cancer (GDSC) and Cancer Cell Line Encyclopedia (CCLE) datasets.
• Applied Pearson correlation for dimensionality reduction and neural network modeling.
• Employed repeated K-fold cross-validation (K=10, 5 repeats) for performance evaluation.

Main Results:

• The NeuPD framework demonstrated superior performance on the GDSC dataset.
• Achieved a Root Mean Square Error (RMSE) of 0.490 and R-squared (R²) of 0.929.
• Outperformed existing drug sensitivity prediction approaches.

Conclusions:

• The NeuPD framework shows significant potential for validating anti-cancer drugs.
• This approach advances personalized medicine by improving drug sensitivity prediction accuracy.
• Further research can leverage NeuPD for broader drug discovery and treatment strategies.