PRODeepSyn: predicting anticancer synergistic drug combinations by embedding cell lines with protein–protein interaction network

Abstract

Although drug combinations in cancer treatment appear to be a promising therapeutic strategy with respect to monotherapy, it is arduous to discover new synergistic drug combinations due to the combinatorial explosion. Deep learning technology holds immense promise for better prediction of in vitro synergistic drug combinations for certain cell lines. In methods applying such technology, omics data are widely adopted to construct cell line features. However, biological network data are rarely considered yet, which is worthy of in-depth study. In this study, we propose a novel deep learning method, termed PRODeepSyn, for predicting anticancer synergistic drug combinations. By leveraging the Graph Convolutional Network, PRODeepSyn integrates the protein–protein interaction (PPI) network with omics data to construct low-dimensional dense embeddings for cell lines. PRODeepSyn then builds a deep neural network with the Batch Normalization mechanism to predict synergy scores using the cell line embeddings and drug features. PRODeepSyn achieves the lowest root mean square error of 15.08 and the highest Pearson correlation coefficient of 0.75, outperforming two deep learning methods and four machine learning methods. On the classification task, PRODeepSyn achieves an area under the receiver operator characteristics curve of 0.90, an area under the precision–recall curve of 0.63 and a Cohen’s Kappa of 0.53. In the ablation study, we find that using the multi-omics data and the integrated PPI network’s information both can improve the prediction results. Additionally, the case study demonstrates the consistency between PRODeepSyn and previous studies.

Pisces: A multi-modal data augmentation approach for drug combination synergy prediction

“…We aim to find the best f θ to enable the predictions on test set { ŷ i } test to approximate the labels { y i } test . Moreover, following the previous works [12, 14–16], we view the prediction probability greater than 0.5 as synergy, otherwise no synergy.…”

“…Positive pairs for contrastive learning We aim to find the best f θ to enable the predictions on test set { ŷi } test to approximate the labels {y i } test . Moreover, following the previous works [12,[14][15][16], we view the prediction probability greater than 0.5 as synergy, otherwise no synergy. Furthermore, for drug representations, we define the SMILES features and molecular graph features of drugs as s and g. Each SMILES string can be represented as s = (s 1 , .…”

“…Likewise, in the stratified CV for cell lines setting, cell lines in the test set have never been seen in the training set. Moreover, PRODeepSyn, GraphSynergy and Pisces share the PPI network which is obtained from the STRING database [50] as in [14].…”

Network approaches for modeling the effect of drugs and diseases

“…PRODeepSyn [ 96 ] is a recently published method that applies GCN to embed cell lines based on omics data and a PPI network and then predicts drug-combination synergy based on the embedding and drug features. A novel idea proposed by this method is to learn a low-dimensional latent representation of cell lines which is based on optimizing a projection to the omics profiles associated with cell lines.…”

Systems pharmacology approaches in herbal medicine research: a brief review

“…By associating targets with biological functions or diseases, an herbcompound-target-function/disease network is constructed and leveraged to study the MOA of an herbal treatment. Proteinprotein interaction (PPI) network could be further employed to interpret the synergistic effect of herbal medicine by analyzing interactions among target proteins (9,10). As this network-based approach is solely based on the network constructed, a significant factor influencing subsequent analysis is the reliability of connections (herb-compound, compound-target, and target-dishttp://bmbreports.org ease) in the network.…”