PANCDR: precise medicine prediction using an adversarial network for cancer drug response

Juyeon Kim,Sung-Hye Park,Hyunju Lee

DOI: https://doi.org/10.1093/bib/bbae088

IF: 9.5

2024-03-17

Briefings in Bioinformatics

Abstract:Pharmacogenomics aims to provide personalized therapy to patients based on their genetic variability. However, accurate prediction of cancer drug response (CDR) is challenging due to genetic heterogeneity. Since clinical data are limited, most studies predicting drug response use preclinical data to train models. However, such models might not be generalizable to external clinical data due to differences between the preclinical and clinical datasets. In this study, a Precision Medicine Prediction using an Adversarial Network for Cancer Drug Response (PANCDR) model is proposed. PANCDR consists of two sub-models, an adversarial model and a CDR prediction model. The adversarial model reduces the gap between the preclinical and clinical datasets, while the CDR prediction model extracts features and predicts responses. PANCDR was trained using both preclinical data and unlabeled clinical data. Subsequently, it was tested on external clinical data, including The Cancer Genome Atlas and brain tumor patients. PANCDR outperformed other machine learning models in predicting external test data. Our results demonstrate the robustness of PANCDR and its potential in precision medicine by recommending patient-specific drug candidates. The PANCDR codes and data are available at https://github.com/DMCB-GIST/PANCDR.

biochemical research methods,mathematical & computational biology

What problem does this paper attempt to address?

The main problem that this paper attempts to solve is to improve the accuracy of Cancer Drug Response (CDR) prediction, especially in the case of limited clinical data. Due to genetic heterogeneity, accurately predicting the drug response of cancer patients is very challenging. Most existing studies have used pre - clinical data (such as cell line data) to train models, but these models have poor generalization ability on external clinical data because of the differences between pre - clinical data and clinical data. To solve this problem, the authors proposed a model named PANCDR (Precision Medicine Prediction using an Adversarial Network for Cancer Drug Response). PANCDR reduces the gap between pre - clinical data and clinical data by using an adversarial network and combines with the CDR prediction model to extract features and predict drug responses. Specifically, the PANCDR model contains two sub - models: 1. **Adversarial Model**: The goal of this model is to distinguish between pre - clinical data and unlabeled clinical data, thereby reducing the distribution differences between the two. 2. **CDR Prediction Model**: This model is responsible for extracting features from the input data and predicting drug responses. The PANCDR model is trained in two steps: - **Step 1**: Fix the weights of the CDR prediction model and train the discriminator of the adversarial model so that it can distinguish between pre - clinical data and clinical data. - **Step 2**: Fix the weights of the discriminator and train the CDR prediction model so that it can predict drug responses and deceive the discriminator. In this way, PANCDR can simultaneously utilize pre - clinical data and unlabeled clinical data during the training process, thereby improving the prediction performance on external clinical data. The experimental results show that PANCDR outperforms other machine - learning methods on external test data (such as the TCGA dataset), demonstrating its potential in precision medicine, especially in recommending patient - specific candidate drugs.

PANCDR: precise medicine prediction using an adversarial network for cancer drug response

DRN-CDR: A cancer drug response prediction model using multi-omics and drug features

TransCDR: a deep learning model for enhancing the generalizability of cancer drug response prediction through transfer learning and multimodal data fusion for drug representation

DeepCDR: a Hybrid Graph Convolutional Network for Predicting Cancer Drug Response

An Ensemble Learning Approach for Cancer Drug Prediction

TCR: A Transformer Based Deep Network for Predicting Cancer Drugs Response

Machine learning based personalized drug response prediction for lung cancer patients

A subcomponent-guided deep learning method for interpretable cancer drug response prediction

GADRP: graph convolutional networks and autoencoders for cancer drug response prediction

Dual-Layer Strengthened Collaborative Topic Regression Modeling for Predicting Drug Sensitivity

An Improved Anticancer Drug-Response Prediction Based on an Ensemble Method Integrating Matrix Completion and Ridge Regression

Abstract 3537: Assessing machine learning models of drug response predictions using orthotopic patient-derived xenograft

DBDNMF: A Dual Branch Deep Neural Matrix Factorization method for drug response prediction

A Novel Approach For Drug Response Prediction In Cancer Cell Lines Via Network Representation Learning

Predicting drug response of tumors from integrated genomic profiles by deep neural networks

GraphCDR: a graph neural network method with contrastive learning for cancer drug response prediction

A programmed cell death-related model based on machine learning for predicting prognosis and immunotherapy responses in patients with lung adenocarcinoma

A novel heterogeneous network-based method for drug response prediction in cancer cell lines

Personalized Anti-Tumor Drug Efficacy Prediction Based on Clinical Data

BANDRP: a bilinear attention network for anti-cancer drug response prediction based on fingerprint and multi-omics

A method combining LDA and neural networks for antitumor drug efficacy prediction