MARSY: a multitask deep-learning framework for prediction of drug combination synergy scores

Abstract

Motivation Combination therapies have emerged as a treatment strategy for cancers to reduce the probability of drug resistance and to improve outcome. Large databases curating the results of many drug screening studies on preclinical cancer cell lines have been developed, capturing the synergistic and antagonistic effects of combination of drugs in different cell lines. However, due to the high cost of drug screening experiments and the sheer size of possible drug combinations, these databases are quite sparse. This necessitates the development of transductive computational models to accurately impute these missing values. Results Here, we developed MARSY, a deep learning multi-task model that incorporates information on gene expression profile of cancer cell lines, as well as the differential expression signature induced by each drug to predict drug-pair synergy scores. By utilizing two encoders to capture the interplay between the drug-pairs, as well as the drug-pairs and cell lines, and by adding auxiliary tasks in the predictor, MARSY learns latent embeddings that improve the prediction performance compared to state-of-the-art and traditional machine learning models. Using MARSY, we then predicted the synergy scores of 133,722 new drug-pair cell line combinations, which we have made available to the community as part of this study. Moreover, we validated various insights obtained from these novel predictions using independent studies, confirming the ability of MARSY in making accurate novel predictions. Availability and implementation An implementation of the algorithms in Python and cleaned input datasets are provided in https://github.com/Emad-COMBINE-lab/MARSY. Supplementary information Online-only supplementary data are available at Bioinformatics online.

CODEX: COunterfactual Deep learning for thein-silicoEXploration of cancer cell line perturbations

“…We evaluated the Zero Interaction Potency (ZIP), which incorporates Loewe Additivity [Loewe, 1953] and the Bliss independence [Bliss, 1939] score. All baselines were extracted from [El Khili et al, 2023].…”

“…The Gene Expression Omnibus accession numbers to obtain the raw datasets are: Srivatsan et al [2020] GSM4150378, Lotfollahi et al [2023] GSE206741, Norman et al [2019] GSE133344. The data from Replogle et al [2022] are available from https://doi.org/10.25452/figshare.plus.20022944 and the prepossessed drug-synergy data from https://github.com/Emad-COMBINE-lab/MARSY/tree/main/data [El Khili et al, 2023]. Further information and code to reproduce the experiments are provided in the CODEX repository at https://github.com/sschrod/CODEX.…”

Building A Foundation Model for Drug Synergy Analysis Powered by Large Language Models

“…For each task, we selected two datasets and two metrics for evaluation. For regression, we included the Pearson Correlation Coefficient (PCC) and Mean Squared Error (MSE) for model evaluation based on datasets D1 [14] and D2 [17]. For classification, we included ROC-AUC (ROCAUC) and Accuracy (ACC) for model evaluation based on datasets D1 and D3 [18].…”

“…For classification, we included ROC-AUC (ROCAUC) and Accuracy (ACC) for model evaluation based on datasets D1 and D3 [18]. These metrics and datasets were widely used in the related work [14, 17, 18, 37, 38]. We selected seven other methods (DeepSynergy, MARSY, TreeComb, SVM [39, 40], TabNet [41], BERT [42] and Lasso [40, 43]) for benchmarking the regression task and seven methods (DeepSynergy, DeepDDs, TreeComb, SVC, TabNet, BERT and Lasso) for benchmarking the classification task.…”

“…Existing drug synergy prediction methods can be broadly classified into two groups. The first group of methods, such as MARSY [17], focus on predicting specific synergy scores, whereas the second group of methods, such as DeepDDs [18], transfer the continuous synergy score into a binary one via thresholding to infer drug combination synergy. However, most existing methods do not incorporate the extensive synergy information from public databases [19] in their predictions.…”