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

Abstract

Drug response prediction in cancer cell lines is of great significance in personalized medicine. In this study, we propose GADRP, a cancer drug response prediction model based on graph convolutional networks (GCNs) and autoencoders (AEs). We first use a stacked deep AE to extract low-dimensional representations from cell line features, and then construct a sparse drug cell line pair (DCP) network incorporating drug, cell line, and DCP similarity information. Later, initial residual and layer attention-based GCN (ILGCN) that can alleviate over-smoothing problem is utilized to learn DCP features. And finally, fully connected network is employed to make prediction. Benchmarking results demonstrate that GADRP can significantly improve prediction performance on all metrics compared with baselines on five datasets. Particularly, experiments of predictions of unknown DCP responses, drug-cancer tissue associations， and drug-pathway associations illustrate the predictive power of GADRP. All results highlight the effectiveness of GADRP in predicting drug responses, and its potential value in guiding anti-cancer drug selection.

MMSyn: A New Multimodal Deep Learning Framework for Enhanced Prediction of Synergistic Drug Combinations

“…24 Compared with PCA, AE methods perform better for learning low dimensional representations of cell lines. 34 Using handcrafted feature methods enhances the interpretability of the models but limits their generalization. Otherwise, data-driven methods can avoid this drawback.…”

“…Most models only use gene expression features of cell lines. ,,, Some models add other omics data, such as somatic mutations and/or copy number variations (CNVs) , or proteomics data. , Omics data are generally high-dimensional, so feature reduction is usually used by selecting specific gene lists ,, or using data-driven dimensional reduction methods, such as principal component analysis (PCA) or AE . Compared with PCA, AE methods perform better for learning low dimensional representations of cell lines . Using handcrafted feature methods enhances the interpretability of the models but limits their generalization.…”

A comprehensive benchmarking of machine learning algorithms and dimensionality reduction methods for drug sensitivity prediction

“…Regarding FE, principal component analysis (PCA) was employed by several contestants of the DREAM7 drug sensitivity prediction challenge [ 39 ], while others utilize autoencoders [ 3 , 25 , 31 , 33 , 34 ] and Wang et al . [ 10 ] employ a matrix factorization approach.…”

FedGTA: Topology-Aware Averaging for Federated Graph Learning

“…Graphs are extensively utilized for modeling complex systems, primarily due to their ability to visually represent the relational information between different entities, which sets them apart from other types of data. Its growing prevalence in recommendation systems [20,53], drug discovery [33,43], and financial risk control [10,40] urges the development of a correspondent graph analysis tool. Graph Neural Networks (GNNs) emerge as a promising approach to achieve state-of-the-art performance in node-level [23,48,58], edge-level [4,6,56], and graph-level [37,44,51] downstream tasks.…”